Columbia Energy Exchange: The Fastest-Growing Line on the Bill Is the One Nobody Reviews

Contents

1. Bills go up because somebody is using more power — a reflex I held for years
2. Twenty-five levers on one bill, and the one swelling fastest gets the least review
3. Ask which part of the bill went up before believing the reason given
4. Worth a look
5. One thing to take away

A state without the means of some change is without the means of its conservation.

—— Edmund Burke, “Reflections on the Revolution in France” (1790)

In the Columbia Energy Exchange episode published on 3 November 2026, host Bill Loveless talks with Katie Dykes, who until this summer was commissioner of Connecticut’s Department of Energy and Environmental Protection; she spent 14 years at that agency and earlier chaired the state’s utility regulatory authority. She says there is no single lever on an electric bill that solves affordability — you have to go looking for twenty-five scattered ones, some at the legislature, some at the utility commission, some at the Federal Energy Regulatory Commission. Her key figure: transmission is still the smallest share of the bill, yet for Connecticut ratepayers that share has nearly doubled since 2015, and a fast-growing part of it is the asset-condition work of replacing ageing poles and wires. Two conditions on all of it: the six New England states share a grid none of them controls alone, and she is describing her own record in office.

Bills go up because somebody is using more power — a reflex I held for years

Before I listened, the model in my head was crude: bills go up because more people are using power. Data centres, AI training, electric vehicles, heat pumps all push demand up, supply fails to keep pace, prices rise. Follow that model and the fix is just as direct — block or slow the data centres, or build generation faster. It is what politicians are shouting in an election year, and I read the news along with it without stopping to doubt.

The same model shapes how I look at holdings: how much will load grow, so how much generation is needed, so which equipment makers book the orders. What I missed is that an electric bill is not one number — it contains generation, transmission, distribution and a set of public-policy programme charges, four blocks each with its own pricing mechanism and its own regulator. I was reading it as a single total, so I could only think of a single cause.

Twenty-five levers on one bill, and the one swelling fastest gets the least review

The thing that woke me up first in this episode is a contrast: New England has no data-centre load growth, and bills there have still run away far enough to become an election issue. Dykes compares ISO New England — the regional body that runs the grid and wholesale market up there — to the PJM everyone is arguing about, multi-state, with different policies in every state; the difference is data centres exist there and not here. Their load growth comes from households switching heating to heat pumps, more electric vehicles on the road, and a resurgence of old-fashioned economic growth in manufacturing.

Her description of load growth is finer than my binary. The extra consumption can raise grid utilisation, spreading the fixed costs of the distribution system over more kilowatt hours so everyone’s unit price looks better; it can also drive costs up. The dividing line is timing — whether grid and generation get built on the schedule that load actually materialises on. And timing is the hardest thing to guess with consumer-driven electrification: nobody knows how fast heat pumps and electric vehicles will go in, and those two loads look different in different seasons, even in different parts of the same day. That pushes her focus onto the distribution system: visibility, control, the ability to dispatch distributed resources, which is where the sweet spot between reliable and affordable sits.

Then the host puts something harder on the table. He cites recent arguments from scholars at Columbia’s Center on Global Energy Policy: we are too quick to blame rising prices on data-centre demand growth, and what they point to is structural — infrastructure planning efficiency, utility incentives, transmission constraints, permitting, and how costs get allocated.

Dykes’s answer is those twenty-five levers. She refuses a single cause, and says every part of the bill has something in it you can improve, with the fixes scattered across different jurisdictions. Then she gives the example that has swelled.

New England has seen rapid growth in utility spending on replacing ageing infrastructure. Transmission is the smallest share of the bill, yet for Connecticut ratepayers it has nearly doubled since 2015, and a large part of that growth is ageing-equipment replacement. She makes her position clear first: poles and wires need fixing and she is the first to say so. The problem is the nature of the category — these asset-condition projects get very little review from FERC, and utilities earn a double-digit return on equity for building them. A process few people question on one side, the best-looking return on the books on the other, both growing out of the same line item.

So the states joined consumer advocates in calling for stronger oversight, and a few weeks ago ISO New England filed a proposal with FERC to tighten review of asset-condition projects. She puts that next to “we are simultaneously arguing for more new transmission”: while you push to build, you have to watch the other costs climbing onto the bill.

The host presses on incentives: are utility regulatory incentives still tilted towards building new things, with no money in getting more out of what exists? She answers flatly — of course. The post-war cost-plus model, giving utilities and shareholders a predictable return for deploying capital into more poles and wires, was the engine of the American economy and is how the grid got built. But today the distribution system has distributed batteries and virtual power plants that can raise utilisation, and they compete directly with the traditional investment case for expanding a substation. Connecticut has attempted performance-based regulation, trying to reward utilities for operating the grid more efficiently. Models are being debated all over the country, she says, and making one work requires a fresh conversation with utility investors about how they get paid.

The sums involved are close to absurd. The host cites research from the nonprofit Power Lines: 51 investor-owned utilities plan to spend 1.4 trillion dollars over the next five years, which that group’s executive director describes as three interstate highway systems and thousands of Hoover dams worth of spending. Dykes’s next line is the hinge of the episode: precisely because so much will be built, the regulatory framework has to be right first. She favours encouraging utilities to right-size projects and build them bigger the first time, since that saves long-run permitting and siting cost; but you only let someone build bigger if you trust they are not overspending in some other category.

State-to-state coordination overturned another assumption of mine. I had assumed six states sharing a grid meant six states tripping each other up. In 2020 they put together a vision statement, all six calling on ISO New England to reform its governance, its transmission planning and its market rules — and getting six states with different policy drivers to sign one document was itself a fight. The result, a few weeks ago, was the six states jointly selecting a 1,200 MW transmission project in northern Maine to reach the relatively low-cost onshore wind up there. States are notoriously bad at agreeing how to split a bill, she says, and this time they agreed even on the cost allocation.

Two more details worth copying down. They reviewed their own renewable portfolio standard and found it was subsidising generation that emits air pollutants and carbon — biomass, landfill methane — out of step with their carbon goals, and reforming it saved customers millions. She also redefined clean energy for me: the New England grid runs heavily on natural gas, and during the war in Ukraine gas prices spiked straight through onto the bill; wind and solar have no fuel cost, so beyond emissions they are a hedge against delivered-fuel price volatility. Once you separate those benefits, the question becomes how to get them at the lowest possible cost, which is where purely regulatory tools like faster permitting and appliance efficiency rules earn their place — LED lighting is one of the great affordability stories, but it is a regulation, the savings are hard to quantify, and that makes it a hard sell today.

ISO New England is also changing how its capacity market credits resources, granting more granular credit based on when each resource can actually deliver. A decade ago the six states were fighting over a rule called the MOPR and whether state investment in clean energy interfered with the capacity market; the new accounting lets people step away from the green-versus-brown argument. Her example stuck with me: offshore wind happens to perform well in exactly the multi-day cold snaps that are the weak point of the New England grid.

The last piece that stopped me is relief customers can feel. Sometimes the most impactful step, she says, is to spend now to avoid spending more later — and some of her biggest affordability wins took the shape of a programme’s cost not getting higher. Then she half-demolishes that herself: such a win does not translate for someone who skipped refilling a prescription or was late on rent because of the utility bill. So they securitised the cost of repairing storm damage on the distribution system, using public finance to spread a lumpy charge over years, and used state bonding to cover the arrearage and hardship programmes that ran up large costs during the pandemic.

Ask which part of the bill went up before believing the reason given

When someone tells me bills are up because of data centres, I now ask which part of the bill went up — a question that swaps a slogan for four checkable categories. New England is a useful control group: no data-centre load growth there, and bills are an election issue anyway, so the demand story explains part of it at most. Demand growth is a real factor; it is incomplete, and even its direction depends on whether buildout timing matches when load arrives.

If you have ever felt that a company says it is investing so things will get better and your bill never does, this episode offers an explanation. Preventive spending cannot produce a visible deduction on a bill; it only keeps a number from getting bigger. Making people see the number come down takes spreading a lumpy cost out or putting public money in front of it, and both of those have short political lifespans.

Looking back at my own holdings, I changed a question. With a regulated company I used to look at how big the capital plan was; now I want to know whether the category it earns best on is also the category with the thinnest oversight. Where those overlap is where the rules get changed next — the ISO New England filing at FERC is exactly that. The question is not limited to utilities; it fits any business that can pass costs through with nobody reviewing the pass-through line by line.

Three things I am unsure about, all of them left there by her. First, she is describing her own tenure: the upside is she knows how every decision came about, the downside is that nobody leads with their worst calls — asked what she would do differently, her answer was to do more. Second, the verb she uses for performance-based regulation is “attempted”, and which states manage to fix the incentives she says remains to be seen; that path requires renegotiating returns with utility investors, and nobody knows whether reform or political patience runs out first. Third, her office cut permitting backlogs by 78%, yet she places the next frontier at public trust and admits she has no answer there: a permitting process asks whether this technology belongs in this place built by this person, while the question in people’s minds is whether a better technology exists and what their community gets out of it, and those two sit outside the process as designed. Speeding a process up assumes the public believes the whole enterprise is worth doing, and what she saw in office was trust in government at a low.

Worth a look

• The Columbia Energy Exchange episode of 3 November 2026, the source for this piece
• Recent analysis from Columbia’s Center on Global Energy Policy on the structural drivers of rising power prices
• ISO New England’s filing at FERC to strengthen review of asset-condition projects
• The 2020 New England states’ vision statement on governance and market reform, and the announcement selecting the northern Maine transmission project
• The nonprofit Power Lines research on five-year capital spending across 51 investor-owned utilities

One thing to take away

The fastest-growing line on a bill is usually the line nobody reviews item by item.

Those two facts growing together is no coincidence. A thinly reviewed category offers good returns and little resistance, so resources flow there; and because it is a small share of the total, everyone’s attention sits on the big blocks instead. Connecticut’s transmission charge is the smallest share of the bill and nearly doubled in a decade, and the part that swelled is precisely the least-reviewed, highest-return class of project.

Something I have tried that works outside investing too. Take one payment you make every month — phone, insurance, building management fees, the sum of your streaming subscriptions — break it into line items, dig out the same statement from three years ago and compare. Then ask one question: who normally looks at the line that grew most? The answer is quite possibly nobody, including you.

Raw Markdown

This article is an educational discussion of investment method. It is not advice to buy or sell any individual security, offers no target prices, and does not analyze any current holding. Investing carries risk; make your own decisions or consult a qualified professional.