Forward vs Historical Test Years for Utilities
A historical test year sets rates on a company's past costs; a forward test year uses its forecast. The gap between them drives regulatory lag and earned ROE.
Educational analysis for professional use. This guide is not investment advice or a recommendation to buy or sell any security, and it is not personalised.
The Test Year a Regulator Picks Decides How Much of a Capex Plan Sits Idle
Every rate case runs off a 12-month test year: the period whose costs, sales volumes and rate base the regulator uses to set the revenue a utility is allowed to collect. Two versions of that period are in use across US jurisdictions. A historical test year looks backward, at a period that has already closed. A forward (or future) test year looks ahead, at a period the utility has not yet completed.
The choice sounds like a filing formality. It is not. A utility on a growing capex plan earns its allowed return only on the rate base that is actually in rates, and the test year decides how quickly new plant gets there. That gap between “put into service” and “earning its allowed return” is regulatory lag, and it is the reason earned ROE trails allowed ROE most sharply in jurisdictions running a historical test year on a fast-growing rate base.
Historical Versus Forward, Defined
A historical test year takes the most recently completed 12 months of actual costs and rate base as filed, sometimes with “known and measurable” adjustments for changes that have already occurred but were not fully reflected in that period, such as a new plant that came online partway through it. The base the regulator approves is anchored to what already happened.
A forward test year takes a projected 12-month period, usually starting some months after the case is filed, and sets rates on the utility’s own forecast of costs, sales and rate base for that period. The base the regulator approves is anchored to what the utility expects to happen, tested and adjusted by the regulator and any intervenors before it is accepted.
Some jurisdictions run a hybrid: a historical base year adjusted forward with pro forma changes that are treated as certain enough to include, without going as far as a full forecast period. That sits between the two and behaves more like a historical test year for anything not on the pro forma list.
How to Tell Which a Jurisdiction Uses
The test-year type is not something a utility chooses freely case by case. It is set by the state’s own procedural rules or statute, and every rate case order states it in its opening sections, usually alongside the exact 12-month window the case is built on. Reading a filing for the phrase “test year” or “test period” near the front of the order, rather than assuming from the company’s general reputation, is the reliable way to find it. A single utility can also file cases in more than one jurisdiction and meet a different test-year rule in each.
What the Choice Does to Earned ROE, Capex Timing and the EPS Build
Allowed ROE is the return a regulator authorises on the equity layer of rate base; earned ROE is what the company actually earns once weather, costs and timing are in. Allowed versus earned ROE covers why the two diverge in general. The test year is the single biggest driver of how wide that gap runs during a capex-heavy period.
Under a forward test year, new plant that the utility forecast correctly is already inside the approved rate base when the case is decided, so it starts earning close to the date it enters service. The lag is small, and mostly a function of how accurate the forecast turned out to be, not of the test-year mechanism itself.
Under a historical test year, any capital placed into service after the test period closes earns nothing until the next rate case resets the base to include it. On a rate base growing at 9-11% a year, that can mean a meaningful share of the year’s new plant sits outside rates for a full case cycle, commonly close to a year, which is the reasoning behind the roughly 12-month lag this site’s rate-base roll-forward uses as a default. The same mechanism flows straight into the EPS build: a utility earning its allowed ROE on a stale, historical rate base reports EPS growth below what the same allowed ROE and rate-base plan would produce with no lag at all.
Riders and Trackers Split the Difference
Most historical-test-year states also allow riders, also called trackers, for specific, defined categories of spending: grid-hardening programmes, environmental compliance, or a named large project. A rider lets that spending earn a return through an incremental rate adjustment outside the full base rate case, on a schedule set by statute rather than by the next full filing.
A rider narrows the lag only for what it covers. The rest of the rate base, everything not named in an approved rider, still waits for the next base case in a historical-test-year jurisdiction. A utility can carry a fast-recovering rider programme and a slow-recovering base rate base inside the same set of accounts, and the two behave differently in an earned-ROE reconciliation even though both sit on the same balance sheet.
A $10bn Rate Base, Worked Two Ways
Take a hypothetical operating company with a $10bn rate base, a 9.5% allowed ROE, and 10% growth in the year, so $1bn of new plant enters service and the rate base ends the year at $11bn.
| Forward test year | Historical test year | |
|---|---|---|
| Rate base earning a return during the year | $11bn (the forecast, already in rates) | $10bn (the closed test period) |
| Return earned on the year’s $1bn growth | $95m (9.5% of $1bn, earned through the year) | $0 (earns nothing until the next case resets the base) |
| Total earnings on rate base for the year | $1,045m | $950m |
Allowed ROE is 9.5% in both columns, so the $95m gap is not coming from the rate itself. It is the value of growth capital that a forward-test-year utility is already earning on, and that a historical-test-year utility only picks up once its next rate case catches the base up. Scale that gap across a multi-year capex plan, and it explains why two utilities on the same allowed ROE and the same rate-base growth rate can post different earned ROE and different EPS growth.
Where the Comparison Goes Wrong
The most common overreach is reading a forward test year as lag-free. It only removes lag to the extent the forecast is accurate: a utility that under-forecasts its own capex still under-recovers in the period, which is why most forward-test-year states run a true-up mechanism.
A closely related error treats “historical test year” as a fixed trait of a company rather than of a jurisdiction. A multi-state utility holding company can run a forward test year in one operating subsidiary’s state and a historical one in another’s, so a consolidated earned-ROE figure blends two different lag regimes without saying so.
A rider does not close the gap company-wide either. It gives one slice of the capex plan its own faster clock; everything outside that named category still runs on whatever test year the base rate case uses.
Regulated Utilities Sector Primer
Rate base, allowed return and the payout are the inputs. This primer takes them through a free-cash-flow-to-equity build, earned net income less the increase in the regulated equity book, to a value per share you can screen against P/E and credit.
The Excel model is the primer's two archetype builds live across 13 sheets: free cash flow to equity, earned net income less the increase in the regulated equity book, discounted at the cost of equity, on rate base, equity ratio and allowed ROE, with the rate-base roll-forward, the FFO-to-debt headroom screen and the funding plan on their own sheets. Change the rate-base growth, the allowed ROE or the cost of equity and the value per share moves; the credit and funding sheets update alongside it.
Frequently Asked Questions
- What is a test year in utility rate regulation?
- A test year is the 12-month period a regulator uses to set a utility's costs, sales and rate base in a rate case. A historical test year uses a past 12 months, sometimes adjusted for changes already known at filing. A forward, or future, test year uses a forecast 12 months instead.
- Why does a forward test year cut regulatory lag compared with a historical one?
- In a forward test year, capital added during the forecast period is already priced into rates when the case is decided, so it starts earning close to the point it enters service. In a historical test year, capital added after the test period closes waits for the next rate case before it earns anything.
- What does a rider or tracker do to the lag?
- A rider or tracker lets a defined category of capex, such as a grid-hardening or environmental programme, get a rate adjustment outside the full base case. It narrows the lag on the spending it covers without changing how the rest of the rate base is set.