The Electric Utility

$15.6 Million Doesn’t Explain What Happened

Today I’m going to talk about the city’s After Action Report: Electric Department Financial Review and Internal Controls Assessment. I found parts of this report frustratingly short on detail. Specifically, while the report quantified the financial problem ($15.6 million) it didn’t offer insight into how this amount broke down between the two factors the report identified: electricity costs and capital expenditures. I think this breakdown is important to understanding the recent financial picture, because it lets you target very specifically when and where the issues occurred.

I decided to try to create this breakdown using the city’s budgets and financial reports. To start, I asked the city to explain how it calculated the $15.6 million figure. They described the methodology as follows:

“The amount of $15.6 million presented in the report to council consists of the budget overruns from FY2025 and FY2026. $15.6 million is based upon $9,365,867 that was advanced to Electric from City’s General Fund and other department reserves in FY2025, $4,070,601 which was the unrestricted cash balance of the Electric Department as of 3/31/2024 (the Electric Department Reserves that were spent in FY2025), and $2,154,534 which was the negative cash balance for the electric department computed on a cash basis from our unaudited financials as of June 10, 2026 for FY2026.”

To translate that into table form, it looks like this:

I don’t believe this is an especially useful way to measure the financial challenge, and it doesn’t really match how the report described this figure, which was “expenditures exceeding budgeted levels.” For one, the Electric Department’s 2025 budget already anticipated the use of $6.5 million in reserves to fill the gap between revenues and expenses. Below is a table showing the budget summary across all funds for FY2025. I added the dotted box to highlight the Electric Department.

As another example, in FY2026 the city raised electricity rates partway through the year, though the adopted budget did not assume that increase. That unbudgeted revenue helped offset the cash deficit. If the deficit is treated as the budget miss, this methodology can understate the expense problem in a given year because unbudgeted revenue offsets the cash deficit.

Further, this methodology only looks at FY2025 and FY2026, even though the After Action Report’s explanation points to capital expenditure budget tracking problems that developed over a longer period.

In other words, $15.6 million represents a cash reconciliation, which conflates a variety of underlying factors. It doesn’t show how much electricity costs or capital expenditures exceeded budget, or when those variances occurred. So next, I tried to do that. The table below summarizes my findings.

I hesitated putting a total on this table. The total isn’t the important part, and I’m not trying to say the city’s figure is wrong because it’s too low. It’s wrong because it mixes together many different concepts that make it less useful as a tool to understand a problem.

My point isn’t to explain a complex problem with a single number. My point is that you can’t do that. The interesting takeaway to me is the breakdown of the total. As I’ll go on to describe, once the pieces are separated, the story becomes clearer: the electricity-cost miss explains itself; the capital expenditure miss does not.

Electricity Cost

This amount was fairly straightforward to measure. I just subtracted budgeted electricity costs in 2025 and 2026 from actual costs in 2025 and 2026 (the latter obtained via public records request).

The After Action Report attributed the budget miss to “capacity market volatility.” Overall, I think that explanation is broadly consistent with the data. There isn’t anything that stands out to me here as anomalous.

Starting at a very high level, electricity costs are volatile. The budgeting history shows that the city generally understood the long-term level of those costs. In aggregate, across the ten years before FY2025, actual electricity expense was, on average, 1% below budget.

But that average hides a lot of year-to-year volatility. The average absolute variance from budget was about 12% during that period.

In that context, misses of 16% in FY2025 and 19% in FY2026 were large, but not outliers.

Let’s look more closely at the detail. Electricity costs include the cost to purchase power and the cost to purchase capacity. The former is relatively easy to understand – it’s the cost the city pays to buy the electricity its customers use.

Capacity purchases are less intuitive. But, simply put, it works like this. The amount of electricity we use on a given day varies considerably over the course of a year (and indeed even within a given day). On the peak demand day (typically in the summer when air conditioning use is the highest) we use about 1.5x-2x more electricity compared to an average day. It’s important to be able to purchase enough electricity to meet the demand on any day (or really, at any given moment), including the peak. For that reason, market regulators require utilities to make sure that enough capacity exists to supply their peak demand (plus a safety margin). The mechanism for ensuring this is the capacity market. Utilities pay power producers just to have enough production capacity on standby to serve their peak demand.

This is a bit like renting a backup power generator. Renting a backup generator has two costs. You pay to have the generator reserved and sitting there in case the power goes out, and you also pay for the fuel if you actually run it. Capacity purchases are like the reservation fee. Power purchases are like buying the fuel.

With that background, let’s look at how the city’s electricity costs trended against two variables: 1) a wholesale power price proxy and 2) capacity market prices¹.

From what I can see, the trends in the city’s electricity costs make sense relative to the trends in the underlying input costs.

I believe the $8.4 million budget miss came from ordinary market volatility. Further, I don’t think the problem is the budget miss itself; electricity costs are volatile. I think the problem is where that volatility sits. I’ve seen the recommendation a few places now (including the After Action Report) that the city should adopt a rate mechanism that passes this volatility through to customers and I agree with that takeaway. The collective balance sheet of the Electric utility’s 10,000 customers has a better capacity to absorb this volatility than the city’s budget. This doesn’t necessarily mean a permanent rate increase. It means rates fluctuate up and down to reflect changes in power costs.

Capital Expenditures

Measuring the budget impact of capital expenditures required a different approach. A one- or two-year budget comparison misses the way capital projects are authorized, delayed, reappropriated, and completed across multiple fiscal years. So I built a capital budget-authority roll-forward beginning in 2018 to determine the cumulative amount of over- or underspending across a longer time-frame.

This analysis starts with 2018 reappropriations (spending that was approved in fiscal years before 2018, but not yet spent), adds new budgeted capital appropriations for each year as they occurred, and subtracts actual capital spending. On this reconstructed basis, cumulative Electric Department capital expenditures exceeded identified capital budget authority by approximately $10.1 million by the end of FY2025 and $11.9 million by FY2026.

One note on what this analysis does and doesn’t capture. This roll-forward includes only capital authority identifiable from adopted budgets, reappropriation schedules, and the separately approved A-Mrazek building acquisition. It does not capture any supplemental appropriations or other adjustments that were not apparent in the annual budgets.

There was also a comment in the After Action Report that says the effect of the capital spending was “to advance five to seven years of planned capital investment into a span of two fiscal years.” I think this is saying that the overspending in recent years didn’t just come from previously-approved work, but also pulled forward projects that were meant to occur in future years. It’s hard to see how this fully reconciles with the data.

The FY2026 five-year capital plan outlined $4.4 million of total capital spending in fiscal years 2027-2030. Most of this ($2.8 million) relates to recurring distribution line replacements, which are not the kind of projects that can easily be pulled forward into a single year. But even if you treat the full $4.4 million as additional budget authority, that still leaves you with $7.5 million in cumulative spending beyond the approved budget authority.

I’ll end with this. If the goal is to measure and identify the underlying causes of the budget problem, the city’s $15.6 million figure is too muddled to do that. A cleaner reconstruction points to roughly $20.3 million of identified budget variance: $8.4 million from electricity costs above budget in FY2025 and FY2026, and $11.9 million from capital spending beyond identified public spending authority.

But those two components should not be treated the same way. The electricity-cost miss is broadly explainable as ordinary-course volatility and acute challenges in the power market. The capital-spending shortfall is different. Capital projects should be authorized, tracked, reappropriated, and visible in the public budget trail. On the record I reviewed, that trail does not fully reconcile. That is why the $11.9 million capital figure is the more important number.

1. This is only a rough approximation of the actual input prices the city incurred. The city’s actual costs depend on things like hedging, power purchase agreements, and bi-lateral capacity contracts. For the wholesale electricity price proxy, I am using EIA’s PJM West wholesale power-price series. This is not the actual price in Kirkwood’s region, but it is a long-running, liquid, downloadable wholesale electricity benchmark.