Post #488: A followup and correction on the 2020 bond issue

Posted on December 10, 2019

For the record, I need to correct what I said about the 2020 bond issue (Post #485).  It’s the part where I summarized it as: “Can they really pay that back?  I think the answer is a qualified yes.”

I believe that’s not true if they borrow the full $35M that they have been authorized to borrow.  I don’t think they can do that without violating their stated minimum reserve requirements for their capital fund, based on their current economic model.  And I didn’t realize that until I saw all the details presented at last night’s Town Council meeting.

I realize it’s too late to do anything, as the Town Council approved the bond issue.  But I think it’s worth stating this for the record.  Just on the off chance that anybody in Town government cares about it.  Presumably, Town staff will redo the reserves calculation just prior to the time of bond issuance using updated information.

My original conclusion was that the Town could afford to issue a $28M bond, without violating their reserve requirement guidelines, but just barely.  Based on their economic assumptions, stretching out the payments as far as possible (e.g., 20 year bond instead of the usual 15 year), burning through maybe $5M in reserves, not borrowing much for the decade or so after this, and suckering NVTA into paying the entire cost of the Patrick Henry garage (when they clearly have no business doing that), then yeah, they could pull this off and not run the reserves down to zero.

My point is, it’s a very near thing.  There’s almost no slack in the system.  They’re maxing out their credit card, so to speak.

But there was still this weird ambiguity about whether they are borrowing $28M or $35M.  The Town’s financial analysis pretty clearly said $28M, and so I assumed they’d dropped the plans to borrow an additional $7M for unnamed contingencies. 

But three things came out at last night’s meeting.  First, that notion that they’ve dropped the additional $7M in bond funding is not right.  I was not the only one confused, as one of the citizen speakers at the public hearing brought it up and asked that this point be clarified.  And the Town is, in fact, keeping open the option to borrow $35M, not $28M.  Second, the acceptable minimum level of reserves in the capital fund was stated as $2M, not zero.  (I have since verified that visually based on the graphs in the prior Capital Improvement Plan.) Third, that graph above — which shows the reserves dropping to $3M — that graph is based on borrowing $28M, not $35M.  At least, I’m pretty sure that’s what they said (and that matches my back-of-the-envelope calculation with Excel, and that matches the figure in the Town’s Capital Improvement Plan document (.pdf)), as shown below.

To be clear, it appears that Town staff talked about borrowing $35M, but modeled borrowing $28M.  And, as I pointed out, they are right at the margins of do-ability, based on their own assumptions, with $28M in borrowings. 

What happens if they borrow the full $35M?  It appears that they never modeled that.

The answer is, reserves drop below $2M sometime around 2025 or so, and they drop below zero by 2027-ish.  So, if I have my facts straight — if the graph above is for a $28M bond issue — then they can’t do a $35M bond issue, under their assumptions, and keep reserves above the $2M threshold.  Or above zero, for that matter.

It’s easy enough to show this.  All you need to know is what the additional payment would be, for a $7M 20-year loan at 3% interest.  (The 3% is their assumption).  And there are at least two ways to get a ballpark figure for that.

Using Excel, I come up with $470,000 per year in additional payments.  (That’s =-PMT(0.03,20,7000000) for you Excel fans out there.)

Or you can just read a number that ought to be close enough, right off the Town’s own bond repayment tables.  Interest rates have been low and stable for the past decade.  In 2012, the Town borrowed $6.3M, but as a 15-year (not 20-year) loan.  So, a little less money, but a somewhat shorter payback period, and roughly the same interest rate that we’re now talking about.  That ought to be in the same ballpark.  So, something in the middle of paying back that 2012 bond.

And that guesstimate gives a little over $500,000 per year in additional payments.

Close enough.

So, if the Town adds $7M to the bond issue, they’ll burn through an additional $1M in reserves (or so) every two years. Ballpark.  Paying that back.

Now take the graph above, and lower the reserve line by $1M in 2022, by $2M in 2024, $3M by 2026, and so on.  That accounts for the burden of paying back that additional $7M — roughly $1M in additional reserve drawn-down for every two years.  I’ve crudely marked that here, in red:

And … by eye, the reserves fall below $2M in 2025, and in fall below zero around 2027 or so.  The additional burden of paying back the $7M extra results in too high a rate of reserve burn to maintain an acceptable reserve level.

So now I think I understand the ambiguity of the analysis.  Again, assuming I haven’t gotten anything wrong, the staff modeled a $28M borrowing, but only talked about a $35M borrowing.  If they’d modeled the $35M borrowing, under the same assumptions, I think they’d have violated their reserve requirements.

What would I need to do to keep reserves above zero, with a $35M 2020 bond issue?  A combination of several events could make that $35M issue feasible without bring reserves below $2M.  The key is that the Town will know a lot more about actual interest rates and tax revenues when they issue the bonds in late 2020, than they do now. 

Possibly, staff anticipate that the interest rate on the bond issue will be well below the modeled 3%/year.  In that case, they could redo the calculation with the actual interest rate, and see that in fact, with the actual rate for the 2020 bond, they won’t violate the minimum reserve requirements even if they borrow the full $35M.  So this may just be Town staff’s way of keeping that $35M option open, even though, formally, based on their model, they should have excluded it.

That said, when I try to back-solve for how low interest rates would have to be to allow that to be true, I come up with pretty close to zero.  That is, when I solve for the interest rate that would result in a $35M 20-year loan at X% yielding the same annual payment as a $28M 20-year loan at 3%, I come up with X<1%.  (Which is an arithmetic expression, not an emoji). Again, using the Excel PMT (loan payment estimator) function.

Now, as a former consultant myself, I probably need to to translate that last paragraph for you.  This financial model is based on assumptions.  And you get to pick your assumptions.  So, if the assumptions you picked yield too low a reserve level, why not just adjust your assumptions (within reason) to get a more acceptable answer.  The little pseudo-imoji above is why.  Possibly, you couldn’t work the numbers on a $35M bond, and maintain the reserve targets, under any plausible interest rate assumption.

Alternatively, you might be able to stay above the $2M reserve threshold if you simply did literally zero non-sewer/water bond financing for the next bond cycle or two.  (I didn’t run the numbers on that.)  But if so, and that’s how they plan to avoid running reserves down to zero in the $35M scenario, that really should have been on the table.  As in, sure, you can borrow the extra $7M now, but if you do, then the money spigot is shut off for a few years while you pay that down.

Yet a third scenario would be increases in meals and lodging tax revenues above the currently-anticipated level.  By the time the issue the bonds, they’ll have had almost a year’s worth of additional information on trend in those tax revenues.  If those revenues are trending above the projected level, that would increase the Town’s ability to fund this out of tax flows, and reduce the burden on reserves.

In short, plausibly, some combination of lower-than-predicted actual interest rates, lower-than-planned future capital spending, or higher-than-projected meals tax revenues could make the funding of the additional $7M feasible.  No one of them might be fully adequate, but some combination might.  In particular, I note that the Town tends to make conservative (high) assumptions on interest rates, and given recent interest rates that Town has paid, and current interest rate trends, I’d be surprised if the Town ended up paying the full 3% rate embedded in the Town’s economic model.

So at that point, as with so much that goes on with the Town of Vienna, I have to throw up my hands and say, I have no clue what the deal is here.  Best guess, Town staff didn’t show the pro-forma reserves calculation for the $35M bond because they didn’t want to rule that out.  By the time the bonds are actually issued, they will have much more accurate information on costs and revenues and can redo the reserves calculation use that updated information.