A rebased price index can look like a price cut when the scale simply changed

A cost forecast can show a dramatic price decline when an imported index series changes from an old reference base to a new one. If the previous observation is 200 and the next is 105 on a different scale, dividing the two directly creates a change that may have nothing to do with prices.

The Bureau of Labor Statistics explains that an index measures movement relative to a reference period and that index-point changes depend on that reference scale. Percentage changes between consistently based observations preserve the underlying movement.

Check the series metadata before interpreting an unusual jump. The index identifier, reference base, observation period and adjustment status belong with the numbers. A column labeled inflation index is not enough to establish that every row is directly comparable.

Assume an invented price index has a value of 200 in the comparison period on its old scale. The publisher rebases that same period to 100 without changing the underlying series. A later observation is 105 on the new scale.

The correct change from the comparison period is 105 divided by 100 minus one, or 5%. On the old scale, the later observation would be 210, giving the same 5% change from 200.

Combining the old 200 with the new 105 produces negative 47.5%. That result is a scale mismatch. It is not evidence that the measured prices nearly halved. The conversion factor is two old-scale points for each new-scale point in this deliberately simple rebasing.

For a pure rebasing to make a chosen reference observation equal 100, divide every observation in the series by that reference observation and multiply by 100. Applying the same factor throughout preserves ratios between periods.

If an earlier old-scale observation was 190, it becomes 95 on the new scale. Its increase to the comparison-period value is 200 divided by 190 minus one, or approximately 5.26%, which is the same as 100 divided by 95 minus one.

Retain the original values and the conversion rule. A finance workbook should make it possible to reconstruct the publisher's series and the normalized analytical view. Overwriting raw observations with manually adjusted numbers can make a later update difficult to reconcile.

An increase from 100 to 105 is five index points and 5%. An increase from 200 to 210 is ten index points and also 5%. The number of points differs because the starting scale differs.

Conversely, adding ten points to an index starting at 100 produces a 10% increase, while adding ten points to an index starting at 200 produces a 5% increase. A rule described as adding ten index points is therefore not equivalent to a 10% escalation.

Use the actual formula in the commercial agreement or approved planning policy. If a report uses percentage changes, label them as percentages. If a contractual mechanism refers to index points, establish exactly which series and base it means before translating the result into a payment.

Suppose an illustrative agreement adjusts a $50,000 charge in direct proportion to the later index divided by the comparison-period index, with no cap, floor or other adjustment. Using 105 divided by 100 gives a revised charge of $52,500.

Using the old-scale pair, 210 divided by 200, gives the same $52,500. Using the mismatched pair, 105 divided by 200, would give $26,250 and create a false reduction of $23,750 from the original charge.

This example isolates the scale issue. Real agreements can specify lags, averaging windows, covered cost shares and replacement-index provisions. Keep those terms intact while correcting the data comparison; rebasing alone does not authorize a change to the contractual mechanism.

Two index series can overlap in time but measure different populations, baskets, locations or price concepts. Making both equal 100 in one common period aligns their displayed levels. It does not make their subsequent movements economically interchangeable.

For example, a consumer-price measure and a supplier-input measure can respond differently to changes in the economy. The BLS distinguishes indexes by their intended coverage. Choosing an appropriate series for a cost assumption requires a separate assessment from normalizing its reference scale.

When a publisher discontinues or materially changes a series, review its methodological guidance and the agreement's replacement provision. A link factor based on one overlap observation is an analytical construction unless the relevant parties or published method support that treatment. Preserve the discontinuity rather than pretending no definition changed.

A monthly observation and an annual average answer different period questions even if they share the same reference base. Comparing a December index with the prior year's annual average can mix a recent point with a full-year average.

Likewise, seasonally adjusted and unadjusted series are not identical versions of the same number. Use the version required by the comparison and preserve its identification. Rebasing does not correct a mismatch in frequency, seasonal treatment or publication vintage.

Where published observations can be revised, save the values used for the approved calculation. A later extraction may differ from the original input. The review should explain whether an updated result comes from a revised observation, a new period or a changed formula.

After converting a purely rebased series, verify that percentage changes across several overlapping periods match the original series within rounding. An unexplained difference suggests a data-selection or conversion problem that should be resolved before the index drives a charge.

Present the reference-scale conversion separately from the economic change. In the example, the level moves from an old 200 to a new 100 because of rebasing, then from 100 to 105 because of the modeled 5% price movement. Keeping those steps visible prevents a formatting change in the data from becoming a fictitious saving in the budget.