Not all federal benefits and tax parameters adjust for inflation the same way — different programs use different price indices, with meaningful differences in how fast they grow. The Consumer Price Index for Urban Consumers (CPI-U) is the most commonly cited measure and is used for federal poverty guidelines and SNAP thresholds. Social Security's COLA uses the CPI-W (urban wage earners), which weights the spending of working-age people more heavily. There's a long-running debate about whether the CPI-E (experimental, elderly-focused) would be more appropriate for Social Security since it weights healthcare — which the elderly spend more on — more heavily. Federal income tax brackets and the standard deduction use the Chained CPI-U (C-CPI-U) since the Tax Cuts and Jobs Act of 2017 — a slower-growing index that accounts for substitution behavior (when prices rise, consumers buy cheaper alternatives), meaning brackets effectively grow more slowly than under the old CPI-U methodology, creating a subtle long-run tax increase embedded in the TCJA. The Average Wage Index (AWI) — which grows faster than inflation — is used for Social Security's taxable earnings base and bend points, meaning Social Security automatically captures more wages as the economy grows. Understanding which index a program uses tells you how its real value changes over time.
Current Law (2026)
Different government programs use different inflation measures to adjust thresholds, benefits, and tax parameters annually.
| Index | Used For | Characteristics |
|---|---|---|
| CPI-U | Federal poverty guidelines, SNAP, many thresholds | Broad urban consumer basket |
| CPI-W | Social Security COLA | Urban wage earners (working-age weighted) |
| CPI-E (experimental) | Proposed for SS COLA | Elderly consumer basket (more healthcare) |
| Chained CPI-U (C-CPI-U) | Tax bracket indexing (since TCJA) | Accounts for substitution effects (grows slower) |
| AWI (Average Wage Index) | SS taxable earnings base, bend points | Wage growth (typically faster than CPI) |
How It Works
The choice of inflation index is not technical minutiae — it's a policy decision that determines who gains and who loses in the federal system, compounding over decades.
CPI-U (Consumer Price Index for All Urban Consumers) is the headline measure most people think of when they hear "inflation." It covers spending patterns of urban consumers broadly, updated through ongoing surveys. It's used for federal poverty guidelines (which trigger eligibility for SNAP, Medicaid, CHIP, and ACA subsidies), for LIHEAP, and historically for tax parameters. CPI-U is an index of fixed basket of goods — it doesn't account for substitution behavior.
Chained CPI-U is the critical difference since the Tax Cuts and Jobs Act of 2017. Chained CPI grows approximately 0.2-0.3% slower per year than CPI-U because it adjusts the consumption basket dynamically — if beef gets expensive and consumers buy more chicken, chained CPI reflects that shift. This sounds technical, but the long-run effect is significant: tax brackets, the standard deduction, IRA contribution limits, HSA limits, FSA limits, AMT exemptions, and most other tax parameters are now indexed using chained CPI. At 0.25% slower growth annually, that compounds to approximately 2.5% cumulative shortfall over 10 years. A tax bracket threshold that would have been $100,000 under CPI-U indexing is roughly $97,500 under chained CPI — meaning more income flows into higher brackets over time. This is sometimes called "stealth bracket creep."
CPI-W (Consumer Price Index for Urban Wage Earners and Clerical Workers) is narrower than CPI-U, covering only households where more than half of income comes from wages. It's used exclusively for Social Security COLAs and to index the SS earnings test threshold. CPI-W is based on working-age consumer spending patterns, which is odd for a program primarily serving retirees. Retirees spend a higher share of income on healthcare and housing — categories with inflation that has consistently exceeded overall CPI. The result: Social Security COLAs may not fully offset the actual loss of purchasing power retirees experience.
CPI-E (experimental) is BLS's estimate of what CPI would look like if weighted to reflect elderly consumer spending. It has historically grown approximately 0.2-0.3% faster than CPI-W annually. Switching Social Security COLAs from CPI-W to CPI-E would increase annual COLA by roughly $50-$100/year for the average beneficiary — meaningful at the margin for beneficiaries with limited other income.
AWI (Average Wage Index) indexes Social Security's taxable earnings base (the wage cap), bend points in the benefit formula, and the SS earnings test. AWI tracks wage growth in the overall economy, which typically outpaces price inflation. This is intentional: it keeps the SS system roughly proportional to economy-wide wages, rather than shrinking in real terms over time. When wages grow faster than prices (as in most of the 2010s and post-pandemic 2022-2023), the SS wage base rises faster than CPI-based thresholds.
How It Affects You
If you pay federal income taxes: Chained CPI indexing is a slow-motion tax increase baked into current law since the TCJA switched from standard CPI-U to the slower-growing Chained CPI-U for tax parameters starting in 2018. Chained CPI grows approximately 0.2–0.3% slower per year than standard CPI-U because it adjusts the consumption basket for substitution behavior — when beef prices rise and consumers shift to chicken, chained CPI reflects that shift. Over 20 years, this compounding produces roughly 4–6% cumulative bracket creep: income that would have stayed comfortably in the 22% bracket under CPI-U indexing flows gradually into the 24% bracket instead. The effect is invisible on any single return — fractions of a dollar per year — but measurable over a decade. In 2026, tax bracket thresholds grew by approximately 2.7% while CPI-U grew roughly 3.0%, widening the gap as the law intends. This chained CPI indexing also applies to the standard deduction, AMT exemption, IRA and 401(k) contribution limits, HSA/FSA limits, estate tax exemption, and most other annually adjusted tax parameters. Reverting to CPI-U would cost an estimated $30-50 billion over 10 years in federal revenue — which is why Congress built this in as a quiet offset in the 2017 TCJA rather than a transparent tax rate increase.
If you receive Social Security: The October COLA announcement — based on third-quarter CPI-W data — determines your January benefit increase. CPI-W covers spending patterns of urban wage earners, not retirees — and retirees typically spend significantly more on healthcare and housing, both of which have historically inflated faster than overall CPI. The 2025 COLA was 2.5%, lower than the prior two years and lower than what the experimental CPI-E index would have produced (CPI-E typically grows 0.2–0.3% faster than CPI-W by reflecting elderly consumer spending patterns). Over a 20-year retirement, the gap between CPI-W COLAs and actual retiree spending growth can erode purchasing power by 10–15% cumulatively. The Medicare Part B premium complicates this further: since Part B is deducted directly from Social Security checks, a large Part B premium increase can effectively negate a positive COLA in a given year. Check the current COLA amount and Part B premium at ssa.gov/cola and cms.gov/newsroom/fact-sheets. Policy proposals to switch SS COLAs from CPI-W to CPI-E have bipartisan support from senior advocacy groups (AARP, Social Security Works) but face estimated costs of $150–200 billion over 10 years — a significant obstacle in any solvency discussion.
If you receive means-tested federal benefits (SNAP, ACA subsidies, Medicaid, CHIP, LIHEAP): The Federal Poverty Guidelines (FPG) — updated each January using CPI-U — are the eligibility thresholds for most of these programs. For 2026, the FPG for a family of 4 is approximately $32,150/year; ACA subsidies phase out at 400% FPG ($128,600 for a family of 4); SNAP eligibility runs to 130% FPG ($41,795 for a family of 4). When your income grows slower than CPI-U, you may become newly eligible for programs you previously didn't qualify for. When your income grows faster than CPI-U — or when your state counts certain income types differently — you may lose eligibility or see reduced benefits. The FPG is updated annually by HHS each January at aspe.hhs.gov/topics/poverty-economic-mobility/poverty-guidelines. For ACA marketplace subsidies specifically, eligibility is based on Modified Adjusted Gross Income (MAGI) compared to FPG — a threshold that changes each year, so recheck your eligibility during each open enrollment period even if your income appears stable.
If you manage retirement plan administration, payroll, or employee benefits: Most key retirement and health savings contribution limits are indexed to Chained CPI-U under IRC § 415(d) and related provisions, updated annually via IRS Revenue Procedures. For 2026, the 401(k)/403(b) elective deferral limit is $23,500; the IRA contribution limit is $7,000 (plus $1,000 catch-up for ages 50+); the HSA annual limit is $4,300/self-only and $8,550/family; the Social Security taxable earnings base (indexed to the faster-growing AWI, not CPI) is $176,100. Because Chained CPI grows slower than standard CPI-U, annual increases in these limits are smaller than overall inflation would suggest — don't assume retirement plan limits will rise by the same percentage as general price increases. The IRS publishes all annually adjusted tax figures in a single annual Revenue Procedure (search "Rev. Proc. 2025" for 2026 parameters at irs.gov/newsroom/irs-announces-pension-plan-limitations) — one document covers 401(k), IRA, HSA, FSA, estate tax, and all bracket parameters, making it the essential annual reference for benefits compliance.
Implementing Regulations
CPI methodology is maintained by the Bureau of Labor Statistics under internal statistical procedures. 29 CFR Part 2510 and related DOL rules reference CPI for benefit indexing. The IRS publishes annual inflation adjustments under IRC § 1(f) through Revenue Procedures (e.g., Rev. Proc. 2025-32 for 2026 parameters) — not CFR. No single CFR section governs indexing methodology across federal programs; each program's authorizing statute specifies which index applies.
Pending Legislation
- Switch SS COLA to CPI-E: Multiple bills in recent Congresses would replace CPI-W with the experimental CPI-E for Social Security COLAs, increasing annual adjustments by roughly 0.2-0.3%/year. Would cost an estimated $150-$200 billion over 10 years. Has bipartisan support from senior advocacy organizations but competes with Social Security solvency concerns.
- Revert tax indexing to CPI-U: The chained CPI switch was a revenue-raiser embedded in TCJA; reverting it would reduce federal revenues by an estimated $30-50 billion over 10 years. No active bill in the 119th Congress has moved on this specifically.
Recent Developments
- Inflation normalizing in 2025-2026, but cumulative price levels remain elevated: CPI data through early 2026 showed overall inflation near the Federal Reserve's 2% target — the lowest in nearly five years — continuing the deceleration from post-pandemic peaks. However, price levels for food, housing, and medical services remain 20-30% above 2020 levels. COLAs based on current low-rate inflation may feel inadequate to retirees whose budgets were set based on pre-2021 prices.
- 2026 food price note: BLS data released April 2026 showed food-at-home prices running approximately 0.6% above year-ago levels — a sharp slowdown from the 10%+ grocery inflation of 2022-2023. This directly affects CPI-U and CPI-W calculations for programs indexed annually.
- Tax bracket growth slowing: Under chained CPI indexing, the 2026 tax bracket thresholds grew by approximately 2.7% over 2025 levels (following the 2025 growth of about 2.8%). For comparison, CPI-U grew roughly 3.0% in the same period — meaning the bracket gap is actively widening, as designed.
The BLS February 2026 CPI release (published March 11, 2026) reported that CPI-U rose 2.4% over the prior 12 months, unchanged from the January 2026 reading, while the Chained CPI-U (C-CPI-U) rose 2.2% over the same period—continuing to run approximately 0.2 percentage points below CPI-U, consistent with the substitution-effect gap described in this article. On a monthly basis, CPI-U rose 0.3% in February (seasonally adjusted), driven primarily by shelter (+0.2%), food (+0.4%), and energy (+0.6%). The CPI-W, which determines Social Security COLAs, rose 2.2% over the 12 months ending February 2026, while medical care services—a category weighted more heavily in the experimental CPI-E—rose 4.1% year-over-year, illustrating the persistent gap between overall inflation and the cost pressures facing retirees. (BLS, March 11, 2026)
The BLS January 2026 CPI release (published February 13, 2026) reported that CPI-U rose 2.4% over the 12 months ending January 2026, down from 2.7% for the 12 months ending December 2025, while the Chained CPI-U (C-CPI-U) rose 2.2% over the same period—continuing to run approximately 0.2 percentage points below CPI-U, consistent with the substitution-effect gap described in this article. On a monthly basis, CPI-U rose 0.2% in January (seasonally adjusted), with shelter (+0.2%) and food (+0.2%) as primary contributors, while energy fell 1.5%. The CPI-W, which determines Social Security COLAs, rose 2.2% over the 12 months ending January 2026, while medical care services rose 3.9% year-over-year—illustrating the persistent gap between overall inflation and the cost pressures facing retirees that underlies the CPI-E debate. (BLS, February 13, 2026)
The BLS December 2025 CPI release (published January 13, 2026) reported that CPI-U rose 2.7% over the 12 months ending December 2025, unchanged from the November 2025 reading, while the Chained CPI-U (C-CPI-U) rose 2.5% over the same period—continuing to run approximately 0.2 percentage points below CPI-U, consistent with the substitution-effect gap described in this article. On a monthly basis, CPI-U rose 0.3% in December (seasonally adjusted), driven primarily by shelter (+0.4%), food (+0.7%), and energy (+0.3%). The CPI-W, which determines Social Security COLAs, rose 2.6% over the 12 months ending December 2025, while medical care services rose 3.5% year-over-year—illustrating the persistent gap between overall inflation and the cost pressures facing retirees that underlies the CPI-E debate. (BLS, January 13, 2026)
The BLS November 2025 CPI release (published December 18, 2025) reported that CPI-U rose 2.7% over the 12 months ending November 2025, down from 3.0% for the 12 months ending September 2025, while the Chained CPI-U (C-CPI-U) rose 2.6% over the same period—continuing to run approximately 0.1–0.2 percentage points below CPI-U, consistent with the substitution-effect gap described in this article. On a monthly basis, CPI-U rose an estimated 0.2% over the two-month period September–November 2025 (BLS did not collect October 2025 data due to a lapse in appropriations), with shelter (+0.2%), energy (+1.1%), and food (+0.1%) as primary contributors over that span. The CPI-W also rose 2.7% over the 12 months ending November 2025, while medical care services rose 3.3% year-over-year—illustrating the persistent gap between overall inflation and the cost pressures facing retirees that underlies the CPI-E debate. Notably, BLS also announced the removal of long-term care insurance from the health insurance index effective with this release, citing market changes that rendered it out of scope for the CPI basket. (BLS, December 18, 2025)
The BLS September 2025 CPI release (published October 24, 2025) reported that CPI-U rose 3.0% over the 12 months ending September 2025, up from 2.9% for the 12 months ending August 2025, while the Chained CPI-U (C-CPI-U) rose 2.9% over the same period—narrowing but still maintaining a gap below CPI-U consistent with the substitution-effect differential described in this article. On a monthly basis, CPI-U rose 0.3% in September (seasonally adjusted), driven primarily by gasoline (+4.1%), food (+0.2%), and shelter (+0.2%), while energy services fell 0.7%. The CPI-W rose 2.9% over the 12 months ending September 2025, and medical care services rose 3.9% year-over-year—continuing to illustrate the gap between overall inflation and the cost pressures facing retirees that underlies the CPI-E debate. (BLS, October 24, 2025)
The BLS August 2025 CPI release (published September 11, 2025) reported that CPI-U rose 2.9% over the 12 months ending August 2025, while on a monthly basis CPI-U rose 0.4% (seasonally adjusted), driven primarily by shelter and food. Core CPI (all items less food and energy) rose 0.3% for the month and 3.1% year-over-year, running above the headline rate and above the Federal Reserve's 2% target. The Chained CPI-U, which governs tax bracket indexing, was not separately reported in this release but continued to track below CPI-U consistent with the substitution-effect gap described in this article. (BLS, September 11, 2025)
The BLS July 2025 CPI release (published August 12, 2025) reported that CPI-U rose 2.7% over the 12 months ending July 2025, while on a monthly basis CPI-U rose 0.2% (seasonally adjusted), driven primarily by shelter. Core CPI (all items less food and energy) rose 0.3% for the month and 3.1% year-over-year, continuing to run above the headline rate and above the Federal Reserve's 2% target. (BLS, August 12, 2025)
The BLS June 2025 CPI release (published July 15, 2025) reported that CPI-U rose 2.7% over the 12 months ending June 2025, while on a monthly basis CPI-U rose 0.3% (seasonally adjusted), driven primarily by shelter. Core CPI (all items less food and energy) rose 0.2% for the month and 2.9% year-over-year, running below the prior month's core rate but still above the Federal Reserve's 2% target. (BLS, July 15, 2025)
The BLS May 2025 CPI release (published June 11, 2025) reported that CPI-U rose 2.4% over the 12 months ending May 2025, while on a monthly basis CPI-U rose 0.1% (seasonally adjusted), driven primarily by shelter. Core CPI (all items less food and energy) rose 0.1% for the month and 2.8% year-over-year, continuing to run above the headline rate and above the Federal Reserve's 2% target. (BLS, June 11, 2025)
The BLS April 2025 CPI release (published May 13, 2025) reported that CPI-U rose 2.3% over the 12 months ending April 2025, while on a monthly basis CPI-U rose 0.2% (seasonally adjusted), driven primarily by shelter. Core CPI (all items less food and energy) rose 0.2% for the month and 2.8% year-over-year, continuing to run above the headline rate and above the Federal Reserve's 2% target. (BLS, May 13, 2025)
The BLS March 2025 CPI release (published April 10, 2025) reported that CPI-U fell 0.1% in March on a seasonally adjusted monthly basis—the first monthly decline in recent months—while rising 2.4% over the 12 months ending March 2025. Core CPI (all items less food and energy) rose 0.1% for the month and 2.8% year-over-year, continuing to run above the headline rate and above the Federal Reserve's 2% target. The monthly decline was driven primarily by falling gasoline prices. (BLS, April 10, 2025)
The BLS February 2025 CPI release (published March 12, 2025) reported that CPI-U rose 2.8% over the 12 months ending February 2025, while on a monthly basis CPI-U rose 0.2% (seasonally adjusted), driven primarily by shelter. Core CPI (all items less food and energy) rose 0.2% for the month and 3.1% year-over-year, continuing to run above the headline rate and above the Federal Reserve's 2% target. (BLS, March 12, 2025)
The BLS Real Earnings release for March 2026 (published April 10, 2026) reported that CPI-U rose 0.9% over the month on a seasonally adjusted basis—the largest single-month increase in recent data—bringing the 12-month CPI-U rate to 3.3% for the period ending March 2026, up from 2.4% for the 12 months ending February 2026. The CPI-W similarly rose 1.1% over the month and 3.3% year-over-year through March 2026, a notable acceleration from the 2.2% annual rate recorded through February 2026. As a result, real average hourly earnings for all employees fell 0.6% over the month, and real average hourly earnings for production and nonsupervisory employees fell 0.9%—meaning nominal wage gains were outpaced by price increases in March. This acceleration in both CPI-U and CPI-W is directly relevant to the trajectory of tax bracket indexing (via Chained CPI-U) and Social Security COLA calculations (via CPI-W). (BLS, April 10, 2026)
The BLS March 2026 CPI release (published April 10, 2026) reported that CPI-U rose 3.3% over the 12 months ending March 2026, up sharply from 2.4% for the 12 months ending February 2026, driven primarily by a 21.2% monthly surge in gasoline prices—the largest single-month gasoline increase since the series began in 1967. On a seasonally adjusted monthly basis, CPI-U rose 0.9%, the largest monthly increase since June 2022. The CPI-W rose 3.3% year-over-year through March 2026, while the Chained CPI-U (C-CPI-U) rose 3.1% over the same period—maintaining a gap of approximately 0.2 percentage points below CPI-U, consistent with the substitution-effect differential described in this article. Medical care services rose 3.7% year-over-year, continuing to outpace overall inflation and illustrating the persistent cost-pressure gap relevant to the CPI-E debate for Social Security COLAs. (BLS, April 10, 2026)