Three-year debt isn’t very relevant to our focus on infrastructure finance. But it may be the simplest way to begin a constructive discussion about whether and/or how to amend the MLF from a lender-of-last-resort to a lender-of actual-loans. Starting with the impact of changing only the rate and leaving the current 3-year max term in place will keep things focused on a few important questions that may ultimately be relevant to federal infrastructure credit facilities.

Three possibilities outline the range: no change (i.e. penalty rates), lower it to about current tax-exempt bond yields for the borrower’s rating or offer a flat Treasury rate to all. Below are some rough estimates (Excel model here):

As suggested in a prior post, JCT and CBO should provide some real estimates for these numbers, however approximate. That’ll help define some questions:

• What do most state & locals need (if anything) that a 3-year facility can help them with? Will a cut from already low 0.60% market rates to 0.12% make that much of a difference? For many borrowers or just a sub-set?

• Anything the MLF offers below the equivalent muni bond rate will likely cause significant market displacement. What’s the real impact of that possibility and who bears the cost? Does this affect the market for those state & locals who don’t utilize a repriced MLF?

• What’s the realistic amount of 3-year loans necessary and useful if the repriced MLF becomes a lender-of-actual-loans?

• If there’s a trade-off between the good a repriced MLF might do for state & locals and market disruption, what’s the scale and how can this be minimized?

Once the MLF moves off penalty pricing, it’s a pretty normal loan program benefit-cost exercise between state & locals, the federal taxpayer and the muni market. The only odd thing is that a lower rate means higher benefits not only to state & locals but also to federal taxpayers due to displacement of muni bonds and reduced tax expenditure cost. So the cost of muni market displacement is probably the binding constraint on BCA optimization here. That makes things interesting on several levels.