Calculate US Inflation Rate by Month Using CPI Formulas
The US inflation rate by month is not the CPI level. It is the percentage change between two CPI index observations. That distinction removes most errors in inflation analysis. A CPI-U reading of 310, 315, or 320 is an index level.
The Bureau of Labor Statistics publishes CPI data monthly, typically around the 10th to 15th day for the preceding month. The market consumes the release as a rate shock: headline month-over-month, core month-over-month, headline year-over-year, core year-over-year. Equity index futures, Treasury yields, breakevens, and dollar pairs reprice on the difference between the published percentage change and the consensus estimate. The arithmetic is simple. The market effect is not.
Understanding the CPI-U: the base series behind the calculation
The Consumer Price Index for All Urban Consumers, CPI-U, is the standard US inflation input used for broad market interpretation. It represents roughly 93% of the US population. It does not measure the exact cost of living for each household. It measures the average change over time in prices paid by urban consumers for a defined market basket.
The reference base is 1982–1984 = 100. That base does not mean prices are “100” in any economic sense. It is an indexing convention. If CPI-U is 300, the index is three times the average level of the 1982–1984 reference period. The inflation rate still requires a comparison against another index level.
The Bureau of Labor Statistics began tracking CPI data in 1913. Current releases are monthly. The mechanical structure is stable:
- CPI-U is the index level for all urban consumers.
- Headline CPI includes food and energy.
- Core CPI excludes food and energy.
- Monthly inflation compares the current CPI level with the prior month.
- Annual inflation compares the current CPI level with the CPI level 12 months earlier.
The calculation uses raw index values. It does not require a model. It does not require seasonal views unless the analyst is comparing seasonally adjusted versus not seasonally adjusted series. It requires only correct pairing of dates.
Inflation is a rate of change. CPI is the level. Confusing the two converts a clean arithmetic problem into a false macro signal.
For cross-asset work, the rate hierarchy matters. A 0.1 percentage-point error in monthly CPI can imply a materially different annualized pace. A monthly change of 0.2% annualizes near 2.4% under simple multiplication and slightly higher under compounding. A monthly change of 0.4% annualizes near 4.9% using compound arithmetic. That spread can move policy-rate expectations by multiple basis points when the release is near a Federal Reserve decision window.
The mechanics of monthly inflation rate calculations
The monthly inflation rate formula is:
((Current Month CPI − Previous Month CPI) / Previous Month CPI) × 100
This is a one-period percentage return. It is identical in form to calculating the percentage move of an equity index from one close to the next. The numerator is the index-point change. The denominator scales the change by the prior index level.
Assume the following CPI-U index levels:
| Month | CPI-U index level | Calculation input |
|---|---|---|
| January | 308.000 | Previous month |
| February | 309.232 | Current month |
The monthly inflation calculation is:
((309.232 − 308.000) / 308.000) × 100
= (1.232 / 308.000) × 100
= 0.400%
The result is a monthly inflation rate of 0.4%. The index rose by 1.232 points. The inflation rate was not 1.232%. The denominator prevents that error.
A second example shows the same structure at a different index level:
| Month | CPI-U index level | Index-point change | Monthly inflation rate |
|---|---|---|---|
| March | 310.500 | — | — |
| April | 310.965 | 0.465 | 0.150% |
| May | 311.742 | 0.777 | 0.250% |
| June | 311.430 | -0.312 | -0.100% |
Each rate uses the immediately preceding month as the base. April is compared with March. May is compared with April. June is compared with May. The base rolls forward every month.
The sequence also shows why index points are not comparable across history without scaling. A 1.0-point increase when CPI is 100 equals 1.0%. A 1.0-point increase when CPI is 300 equals 0.333%. The same index-point move has one-third the inflation rate at the higher index level.
A robust monthly calculation process is narrow:
1. Select the CPI series. CPI-U for broad headline analysis. Core CPI if excluding food and energy.
2. Confirm whether the index is seasonally adjusted or not seasonally adjusted. Do not mix them.
3. Take the current month’s index level.
4. Take the prior month’s index level from the same series.
5. Subtract prior from current.
6. Divide by the prior month’s index level.
7. Multiply by 100.
8. Round only after the calculation, not before it.
Rounding before calculation can alter the rate by several basis points when the index move is small. This matters when the market is trading a 0.2% versus 0.3% month-over-month print. CPI releases are often interpreted at one decimal place, but the internal arithmetic should use more precision where available.
Determining annual trends: the 12-month comparison formula
The 12-month inflation rate uses the same percentage-change structure. The comparison window changes.
((Current Month CPI − CPI from 12 Months Ago) / CPI from 12 Months Ago) × 100
If the current index is June 2025, the base is June 2024. Not December. Not January. Not an average of the prior year. The formula is a point-to-point 12-month comparison.
Example:
| Month | CPI-U index level |
|---|---|
| June, prior year | 300.000 |
| June, current year | 309.000 |
Calculation:
((309.000 − 300.000) / 300.000) × 100
= (9.000 / 300.000) × 100
= 3.000%
The annual inflation rate is 3.0%. The CPI index is 309.000. The two numbers describe different objects.
The 12-month rate is sensitive to base effects. If the CPI index rose sharply in the comparison month one year earlier, the current annual rate can fall even when the latest monthly rate is positive. If the base month was weak, the annual rate can rise despite a moderate current monthly print.
This is not statistical noise. It is embedded in the formula. The denominator and starting point define the year-over-year rate.
Consider a simplified chain:
| Month | CPI-U level | Monthly change | 12-month effect |
|---|---|---|---|
| Month 0 | 300.000 | — | Base month |
| Month 1 | 303.000 | 1.000% | Large early increase |
| Month 12 | 309.000 | — | Current comparison to Month 0 |
| Month 13 | 309.618 | 0.200% | Comparison shifts to Month 1 |
At Month 12, the 12-month inflation rate versus Month 0 is 3.0%. At Month 13, the current CPI is higher than Month 12 by 0.2%. But the year-over-year base has shifted from 300.000 to 303.000. The 12-month rate becomes:
((309.618 − 303.000) / 303.000) × 100
= 2.184%
Annual inflation falls while monthly inflation is positive. That is a base effect, not a contradiction.
For stock index analysis, this is critical. Equity markets do not only price the published annual rate. They price the marginal information in the latest monthly run rate. A year-over-year deceleration caused by base effects can coexist with an uncomfortable month-over-month trend. The rates market usually detects that distinction faster than index-level commentary.
Distinguishing index levels from percentage changes
Most CPI errors are category errors. They treat a level as a rate, or a rate as a level. Market analysis requires a fixed taxonomy.
| Item | What it is | Example | Market use |
|---|---|---|---|
| CPI-U index level | Point-in-time price index | 312.500 | Raw input |
| Monthly inflation rate | Percent change from prior month | 0.3% | Short-run inflation momentum |
| 12-month inflation rate | Percent change from same month one year earlier | 3.2% | Annual inflation gauge |
| Core CPI | CPI excluding food and energy | Index or rate | Underlying trend proxy |
| Headline CPI | CPI including all major categories | Index or rate | Household-facing inflation measure |
A CPI-U index at 312.500 cannot be compared directly with an inflation target of 2%. The 2% target is a rate concept. The index is a level concept. Only the percentage change in the index can be evaluated against inflation-rate benchmarks.
The same distinction applies across components. Medical care services, shelter, apparel, transportation services, and food away from home each enter the CPI structure through price indexes and relative weights. Some categories contain highly specific consumer services, including dental services, where separate operational domains such as oral hygiene and dental care may explain price behavior at the household level but do not replace the CPI methodology. The CPI framework aggregates observed price changes into a standardized index. It is not an invoice-level ledger.
For market work, four rate forms are commonly separated:
- Month-over-month headline CPI. The immediate all-items inflation signal. Sensitive to food and energy.
- Month-over-month core CPI. The cleaner short-run signal for underlying services and goods pressure.
- Year-over-year headline CPI. The broad public inflation reference. Sensitive to base effects.
- Year-over-year core CPI. The slower trend indicator watched for policy persistence.
The monthly core rate often has the highest signal value when the market is repricing the expected path of policy rates. The year-over-year headline rate often has the highest media visibility. These are not the same function.
Annualizing monthly CPI: useful, but easy to overstate
Monthly inflation can be annualized. That does not make it a forecast. It converts a one-month pace into a 12-month equivalent.
There are two common methods.
| Monthly CPI change | Simple annualization | Compound annualization |
|---|---|---|
| 0.1% | 1.2% | 1.21% |
| 0.2% | 2.4% | 2.43% |
| 0.3% | 3.6% | 3.66% |
| 0.4% | 4.8% | 4.91% |
| 0.5% | 6.0% | 6.17% |
Simple annualization multiplies the monthly rate by 12. Compound annualization applies the monthly rate for 12 consecutive months:
((1 + monthly rate)¹² − 1) × 100
Use decimal form inside the formula. A 0.3% monthly rate is 0.003, not 0.3.
For example:
((1 + 0.003)¹² − 1) × 100
= 3.66%
Annualization is useful for regime classification. It is not a substitute for the actual 12-month CPI formula. A single 0.4% monthly print does not establish a 4.9% inflation year. It states that the one-month pace, if repeated for 12 months, compounds to roughly 4.9%.
The distinction matters in equity index reactions. A high monthly print after several soft months has a different implication than a high monthly print inside a six-month sequence of firm readings. The market response depends on distribution, not only one observation.
A simple inflation-momentum grid is often more informative than one annualized number:
| Last three monthly CPI readings | Three-month pattern | Market interpretation |
|---|---|---|
| 0.1%, 0.1%, 0.2% | Low and stable | Disinflation signal intact |
| 0.2%, 0.3%, 0.3% | Moderate firming | Policy-cut probability declines |
| 0.4%, 0.4%, 0.3% | Persistent high run rate | Front-end yields usually reprice higher |
| -0.1%, 0.0%, 0.2% | Rebound from weak base | Requires component check |
| 0.5%, 0.1%, 0.1% | One-month outlier | Lower persistence unless components repeat |
The grid is not predictive certainty. It is a classification of current inflation momentum. Mean reversion is common in volatile components. Persistence is more relevant in shelter and services categories.
Core CPI vs. headline inflation: filtering market volatility
Core CPI excludes food and energy prices. The purpose is not to declare food and energy irrelevant. The purpose is to reduce short-horizon volatility and isolate longer-term inflation pressure.
Headline CPI can move sharply because of gasoline, electricity, utility gas service, or food categories. Core CPI removes food and energy, but it still includes shelter, medical care, transportation services, recreation, education, apparel, and other goods and services. Core is not a complete inflation measure. It is a filtered one.
The difference between headline and core matters by asset class.
| CPI measure | Includes food and energy | Typical volatility | Policy signal | Equity-index relevance |
|---|---|---|---|---|
| Headline CPI | Yes | Higher | Medium | Strong through consumer income and sentiment |
| Core CPI | No | Lower | Higher | Strong through rates, discount factors, and margins |
| Monthly headline | Yes | High | Medium | Immediate index-futures reaction if surprise is large |
| Monthly core | No | Medium | High | Strong signal for Fed path repricing |
| 12-month headline | Yes | Medium | Medium | Public inflation perception |
| 12-month core | No | Lower | High | Trend persistence and valuation pressure |
Core CPI usually receives more weight in policy-rate expectations because central banks attempt to distinguish persistent inflation from relative-price shocks. A gasoline move can reverse. Services inflation can persist through wages, rents, and contracts. That persistence has a different duration profile.
For global stock indexes, the transmission chain is mechanical:
1. CPI surprise changes expected policy-rate path.
2. Expected policy-rate path changes Treasury yields.
3. Treasury yields change equity discount rates.
4. Discount-rate shifts affect duration-sensitive sectors first.
5. Currency and commodity channels transmit the move globally.
6. Non-US indexes reprice through dollar funding, export sensitivity, and local rate expectations.
A 10-basis-point move in two-year Treasury yields after CPI can carry more information than the index-point move in the S&P 500 during the first minute. Equity futures often react first. Rates usually define whether the move persists beyond the initial liquidity window.
The CPI release is not one number. It is a matrix: headline, core, monthly, annual, level, base effect, and surprise versus consensus.
Building a clean monthly US inflation table
A correct inflation table has one row per month and separates levels from rates. It should not overwrite the raw CPI series. It should calculate rates in adjacent fields.
A minimal structure:
| Date | CPI-U level | Monthly inflation | CPI-U level 12 months prior | 12-month inflation |
|---|---|---|---|---|
| Month A | 300.000 | — | — | — |
| Month B | 300.600 | 0.200% | — | — |
| Month C | 301.502 | 0.300% | — | — |
| Month M | 309.000 | 0.250% | 300.000 | 3.000% |
The first monthly value requires a prior month. The first 12-month value requires a value from 12 months earlier. Blank cells are not errors. They reflect insufficient lookback.
The calculation should preserve series integrity:
- Do not mix CPI-U with CPI-W.
- Do not mix seasonally adjusted and not seasonally adjusted index levels.
- Do not calculate a year-over-year rate from an average unless the target metric is explicitly an average.
- Do not treat a rounded published percentage as the raw input for further rate calculations.
- Do not compare an index level with a percentage target.
- Do not infer a future inflation rate from the formula. The formula calculates observed change only.
Spreadsheet implementation is direct. If current CPI is in cell B14 and previous-month CPI is in B13, monthly inflation is:
(B14 − B13) / B13
Format the result as a percentage. If current CPI is in B14 and the CPI from 12 months prior is in B2, the 12-month rate is:
(B14 − B2) / B2
Again, format as a percentage. The multiplication by 100 is implicit when the spreadsheet percentage format is used. It is explicit when writing the formula manually.
The main operational risk is not formula complexity. It is reference error. A one-row shift converts the monthly rate into a false series. A mismatch between seasonally adjusted and not seasonally adjusted observations creates a contaminated signal. A copied formula that references the wrong base month can produce a plausible but incorrect number.
Reading CPI through market probability, not direction
For stock-market analysis, CPI is a probability input. It is not a deterministic equity signal. A lower-than-expected CPI print can raise indexes if yields decline. It can also coincide with weaker cyclicals if the release is interpreted alongside deteriorating growth data. A higher-than-expected CPI print can pressure valuation multiples. It can also support nominal revenue expectations in selected sectors. The sign depends on the macro mix.
The cleaner interpretation starts with the surprise:
| CPI outcome versus consensus | Rates-market impulse | Equity-index bias | Required confirmation |
|---|---|---|---|
| Core monthly below consensus | Lower front-end yields | Positive for duration equities | Breadth and real-yield response |
| Core monthly at consensus | Limited repricing | Positioning-dependent | Component detail |
| Core monthly above consensus | Higher front-end yields | Negative for long-duration equities | Persistence across services |
| Headline high, core stable | Mixed | Sector dispersion | Energy and food contribution |
| Headline low, core firm | Mixed to negative | Rates dominate if core surprise is material | Core services detail |
The market does not price CPI in isolation. It prices CPI relative to expectations, Federal Reserve reaction function, labor-market data, wage growth, retail sales, PMI surveys, and financial conditions. CPI is a high-weight node in that system because it changes the distribution of future policy rates.
The arithmetic still comes first. If the monthly rate is miscalculated, the downstream market interpretation is invalid. If the 12-month rate is confused with annualized monthly inflation, the policy conclusion is unstable. If core and headline are merged into a single inflation label, the signal loses resolution.
Final technical pivot
Calculating the US inflation rate by month requires one formula: current CPI minus prior CPI, divided by prior CPI, multiplied by 100. Calculating the annual rate requires the same formula with the CPI level from 12 months earlier as the base. CPI-U is the level. Inflation is the percentage change. Core CPI is the filtered series excluding food and energy. Headline CPI is the all-items series.
For market use, the pivot is narrow. A monthly core CPI pace at or below 0.2% is broadly consistent with a lower inflation run rate if sustained. A 0.3% pace is borderline. A 0.4% pace or higher raises the probability of tighter rate-path repricing, especially if repeated across consecutive releases. The probability shift is not in the index level. It is in the rate of change.