Break-Even, Payback and Comparing Costs Over Time

Payback is the easiest measure and the weakest

Simple payback divides the extra cost of the better option by the saving it produces each year, and returns the number of years until the two are level. It is popular because it needs two numbers and no assumptions, and it is popular for the same reason it is weak.

It ignores everything that happens AFTER the payback date. Two measures with a four-year payback are treated as identical whether one lasts five years and the other twenty-five, which is precisely the distinction that decides which is worth doing. Used as a screening filter it is fine; used as a decision rule it systematically rejects long-lived investments in favour of short-lived ones.

It also ignores the time value of money, which flatters long paybacks. A saving arriving in year twelve is worth substantially less than the same nominal sum arriving next year, and a payback figure treats them as equal. The correction is to discount the cash flows, and the effect is largest exactly where the simple figure is already marginal.

Different lifespans: annualise, or the comparison is meaningless

A floor covering that costs less and lasts ten years and one that costs more and lasts twenty-five cannot be compared on their purchase price, and they cannot be compared on their total cost either, because the two totals buy different amounts of time.

The correct comparison converts each option into an EQUIVALENT ANNUAL COST: the level annual payment that has the same present value as the option over its own life. That puts a ten-year option and a twenty-five-year option on the same footing without inventing a common horizon, and it is the reason the flooring comparison here returns an annualised figure rather than a total.

The alternative, and the one used in whole-life studies of buildings, is to set a common study period and include the replacements and the residual value that fall inside it. Both methods agree; the annualised form is simply easier to read when the question is which of two products to buy.

Getting this wrong has a direction. Comparing on total cost favours whichever option wears out sooner, because its total covers fewer years. That bias is why cheap short-life specifications win comparisons that a whole-life analysis would reverse.

The discount rate is a choice, and it changes the answer

Discounting expresses the fact that money available now is worth more than money available later, because it can be used in the meantime. The rate is not a market observable — it is a judgement about what the money would otherwise do.

For a household the honest rate is usually the cost of the alternative: the mortgage rate if the work would be borrowed, the savings rate if it would come from savings. For a public body it is a published rate set by treasury guidance. For a business it is the cost of capital. Different payers can correctly reach different conclusions about the same investment, and neither is making an error.

The sensitivity is real. At five per cent, a pound twenty years out is worth about thirty-eight pence today; at ten per cent, about fifteen. Any analysis whose conclusion flips between plausible rates has not produced a decision, it has produced a dependency — and saying which input the answer turns on is more useful than reporting the answer.

One rule must not be broken: nominal cash flows are discounted at a nominal rate, and real — inflation-adjusted — cash flows at a real rate. Mixing them, by projecting energy prices upward with inflation and then discounting at a real rate, double-counts inflation and is the most common technical error in these models.

Escalation: a saving that grows is worth more than one that does not

A measure that saves a fixed quantity of energy saves a rising amount of MONEY if energy prices rise. That makes escalation a genuine part of the analysis rather than a refinement, and it is why energy-saving investments look better over a long horizon than a first-year saving suggests.

It is also the least reliable input in any of these calculations. Energy and material prices are volatile, policy-driven and not forecastable with any confidence over the horizons involved, and an escalation assumption compounds — so a small difference in the rate becomes a large difference in year twenty.

The defensible treatment is to state the assumption, to run the analysis at more than one rate, and to present the conclusion as a range. A single payback figure computed on a single escalation assumption is a precise answer to a question nobody can answer precisely.

Price escalation applies just as much backwards. A rate, a replacement cost or an insurance valuation from a past date has to be brought forward by an index to mean anything today — which is what an escalation calculator does, and why the index chosen matters as much as the arithmetic.

The horizon that matters is usually the owner's, not the asset's

A break-even point is only relevant if the decision-maker is still there to reach it. A refinance that breaks even in four years is a poor decision for someone moving in two, regardless of how sound the arithmetic is, and the same applies to an insulation measure, a boiler, or a solar array on a house that is about to be sold.

That makes the holding period an input rather than a detail, and it is the input most often left unstated. Two people can receive identical quotes, run identical calculations and correctly make opposite decisions.

Rent-against-buy is the extreme case, and the one where a simple comparison misleads most. The transaction costs of buying and selling are large and concentrated at the ends, so the break-even horizon is typically measured in years rather than months; the outcome depends on house price growth, which is unknowable; and it depends on what the deposit would have earned elsewhere, which is an opportunity cost that never appears on a bank statement. A calculator can compare the monthly cash flows honestly. It cannot tell anyone whether to buy a house.

Resale recovery is not a return

Renovation payback is routinely quoted as a percentage of cost recouped at resale, and the figures are almost always below one hundred per cent. It is worth being plain about what that means: a project recovering seventy per cent of its cost has LOST thirty per cent in money terms. It is not an investment with a low return; it is a purchase with a partial refund.

That is not an argument against doing it. Most renovation is bought for the use of the space over the years of living in it, and the recovery figure is a measure of how much of the cost the market gives back at the end rather than a measure of whether it was worth doing. The mistake is only in calling it a return.

The figures also vary by market and by work type in a consistent pattern: modest repairs and presentation work recover a higher proportion than large discretionary projects, and work that brings a property in line with its neighbourhood recovers more than work that takes it beyond. A national average applied to a specific house is one of the weakest inputs on this site, and the pages say so.

What these models leave out

Every comparison here counts the cash flows it was given. Maintenance, cleaning, downtime, the cost of replacing something disruptive to replace, residual value at the end of the study period, and the risk that an option fails early are all real and are all outside the arithmetic unless they were entered.

Some of the omissions are not financial at all. Comfort, noise, appearance, resilience, and the value of not having the work done again for twenty years do not convert into a discount rate, and an analysis that ignores them is answering a narrower question than the one being asked.

So the honest output of a life-cycle comparison is not a winner. It is a ranking with the assumptions written beside it and a note of which assumption the ranking depends on — and where two options come out within a few per cent of each other, the correct conclusion is that the financial analysis does not decide it.