Bureau of Labor Statistics — CPI, Employment & Economic Indicators

The Bureau of Labor Statistics is the most market-moving statistical agency in the world. A single monthly release — the Employment Situation report, published at 8:30 a.m. on the first Friday of each month — can shift 10-year Treasury yields by 10 or more basis points within milliseconds of publication, move equity markets by 1–2%, and reprice Federal Reserve rate expectations by tens of basis points before the opening bell. BLS is the principal federal statistical agency for labor economics, established under 29 U.S.C. § 1 within the Department of Labor but operationally independent: by longstanding norm, the Commissioner is Senate-confirmed and statistical releases cannot be previewed or altered by political appointees. With roughly 2,400 staff and a budget of ~$700 million annually, BLS produces the price, employment, pay, productivity, and safety data that anchor U.S. economic policy, collective bargaining, Social Security cost-of-living adjustments, and trillions of dollars in indexed financial contracts.

Legal Authority

BLS operates under 29 U.S.C. § 1, which directs the Commissioner of Labor Statistics to "acquire and diffuse … useful information on subjects connected with labor." The Commissioner is presidentially appointed and Senate-confirmed under 29 U.S.C. § 2. BLS statistical practices are governed by OMB Statistical Policy Directive No. 1 (SPD 1), which requires that federal statistical agencies operate free from political interference and that no agency personnel preview or alter data before public release. The Confidential Information Protection and Statistical Efficiency Act (CIPSEA, 44 U.S.C. § 3572) protects the identity of BLS survey respondents and prohibits disclosure of individually identifiable data for non-statistical purposes. BLS data collection authority flows from the Paperwork Reduction Act (44 U.S.C. § 3501), which requires OMB clearance for all federal surveys; BLS maintains OMB approvals for each of its major surveys (CPS, CES, CPI, JOLTS, ECI, OEWS, SOII). The Federal Statistical System is coordinated by the Chief Statistician in OMB's Statistical and Science Policy Office, which sets government-wide standards that BLS must follow.

Key Mechanics

BLS operates two parallel data-collection pipelines — establishment surveys and household surveys — that produce distinct but complementary labor market pictures:

Establishment surveys (payroll-based): The Current Employment Statistics (CES) survey collects payroll records from ~145,000 business and government establishments. Data are submitted electronically via electronic data interchange (EDI), web reporting, or Touchtone Data Entry. Because response is voluntary but employers participate at high rates (~74% in time for initial release), BLS uses a "weighted link relative" technique to estimate counts for non-respondents and then benchmarks annually against the universe of employers covered by state Unemployment Insurance (UI) tax records — a near-census of U.S. employment. This benchmark revision, published each March, can significantly revise prior-year payroll estimates.

Household survey (CPS): The Current Population Survey interviews a rotating sample of ~60,000 households across ~800 Census Primary Sampling Units. Each household participates for four consecutive months, exits for eight months, then re-enters for four more. Interviewers use Computer-Assisted Telephone Interviewing (CATI) and Computer-Assisted Personal Interviewing (CAPI). The CPS is conducted by the Census Bureau under contract to BLS. The unemployment rate is derived from the CPS reference week (week containing the 12th of the month); respondents are classified into employment, unemployment, or not-in-labor-force categories based on activity during that week.

Price collection (CPI): BLS field economists visit ~23,000 retail outlets monthly across 75 urban areas to collect ~80,000 price quotes. Outlets are selected via probability-proportional-to-size sampling from a Point-of-Purchase Survey. Items within outlets are sampled using a similar probability framework. New outlets and items are rotated in continuously to keep the sample current. Prices are aggregated using a geometric mean formula at the elementary (item-area) level, then combined using Laspeyres weights derived from the Consumer Expenditure Survey (CEX).

Release security: BLS operates a strict lockup system for its most market-sensitive releases. Credentialed journalists enter a physical or virtual lockup 30 minutes before release and may prepare stories but cannot transmit until 8:30 a.m. ET. BLS's own network connections are severed during lockup and restored at exactly 8:30:00. This simultaneous-release model is enforced by SPD 1 and is essential to maintaining market integrity.

What It Produces

Employment Situation (Jobs Report)

The Employment Situation, released on the first Friday of each month, is simultaneously the most anticipated and most revised economic data release in the world. It draws from two entirely separate surveys:

The Current Employment Statistics (CES) survey — a payroll survey of approximately 145,000 business establishments and government agencies, representing roughly 697,000 individual worksites — produces the headline nonfarm payroll number (jobs added or lost in the prior month) plus breakdowns by industry, hours, and hourly earnings. Because the CES is a survey of employers, it counts jobs, not workers (one person working two jobs appears twice). The initial estimate is published with just ~74% of eventual sample responses in hand; it is then revised twice (preliminary revision the next month; final revision the month after) and again in annual benchmarking against state unemployment insurance records — meaning the "jobs number" the market reacts to can be revised substantially. The March 2024 annual benchmark revision, for example, reduced the prior year's job count by 818,000.

The Current Population Survey (CPS) — a household survey of ~60,000 households conducted jointly with the Census Bureau — produces the unemployment rate (U-3) and a range of broader labor force measures. The headline U-3 rate counts people without jobs who are available for work and actively searched in the prior four weeks. BLS also publishes U-6, the broadest measure, which includes discouraged workers (gave up searching), marginally attached workers (want work but haven't searched recently), and those working part-time for economic reasons — U-6 typically runs 5–7 points above U-3.

Consumer Price Index (CPI)

CPI is the most politically charged statistical product in the federal government — it determines the annual cost-of-living adjustment (COLA) for 70 million Social Security beneficiaries (via CPI-W), adjusts roughly $1 trillion in indexed Treasury securities (TIPS), sets automatic adjustments in millions of private contracts, and drives the Federal Reserve's inflation narrative even though the Fed officially targets PCE (see BEA). BLS collects approximately 80,000 price quotes monthly from roughly 23,000 retail outlets across 75 urban areas, covering a market basket of goods and services derived from the Consumer Expenditure Survey.

Key methodological points:

• CPI-U (all urban consumers) is the headline measure, covering 93% of the U.S. population; CPI-W (urban wage earners) is used for Social Security COLAs and represents ~32% of the population.
• Core CPI (all items excluding food and energy) is the Fed's secondary inflation focus — food and energy are excluded as too volatile for detecting underlying inflation trends.
• Supercore (services excluding shelter) emerged as an additional Fed focus in 2022 because it better captures wage-driven service inflation.
• The shelter problem: Owners' Equivalent Rent (OER) — the estimated rent homeowners would pay to rent their own homes — makes up roughly 34% of core CPI. OER is calculated from a rolling survey that lags actual market rent changes by 12–18 months, creating a persistent wedge between observed housing costs and measured inflation. From 2022–2024, this lag caused CPI to overstate then understate housing inflation relative to actual market conditions, complicating Fed policy.
• Laspeyres vs. Fisher index: CPI uses a fixed-weight Laspeyres formula, which holds the market basket constant and therefore overstates inflation by not capturing consumer substitution toward cheaper alternatives. PCE (BEA) uses a chain-weighted Fisher index that corrects for substitution — this is the primary reason CPI systematically runs 0.3–0.5 percentage points above PCE on an annual basis.

Producer Price Index (PPI)

PPI measures price changes from the seller's perspective — what producers receive for their output rather than what consumers pay. Because input-cost changes typically flow through to consumer prices with a lag, PPI often leads CPI as an inflation signal. PPI is also used in escalation clauses in government contracts (particularly defense and construction) and in business-to-business pricing agreements.

JOLTS (Job Openings and Labor Turnover Survey)

The Job Openings and Labor Turnover Survey, published monthly, tracks the number of unfilled job openings (a demand signal), hires, and separations (quits + layoffs + discharges). The quit rate — the share of workers voluntarily leaving their jobs — is the Federal Reserve's preferred measure of worker confidence in labor market conditions; a high quit rate signals workers feel they can easily find better jobs. The Fed under Chair Powell frequently cited JOLTS during 2022–2024 as evidence of labor market tightness. Data covers approximately 20,000 business establishments.

Employment Cost Index (ECI)

The Employment Cost Index, released quarterly, is the Federal Reserve's preferred measure of wage growth because it controls for compositional shifts in the workforce — unlike average hourly earnings from the jobs report, which can change simply because the mix of jobs (high-wage vs. low-wage) shifts. ECI measures the cost to employers of a fixed bundle of labor (wages plus benefits) for a fixed set of occupations. Fed communications frequently cite ECI alongside PCE as the key inflation inputs.

Other Major Products

• Occupational Outlook Handbook (OOH): biennial publication covering ~900 occupations; projections for 10-year employment growth, typical wages, and educational requirements; widely used in workforce development and career counseling.
• Occupational Employment and Wage Statistics (OEWS): annual wage survey covering 800+ occupations in 600+ metro areas; the primary data source for prevailing wage determinations used in H-1B, H-2A, and Davis-Bacon wage decisions.
• Productivity and Costs: quarterly measures of output per hour in the nonfarm business sector; labor productivity growth is a fundamental driver of non-inflationary wage growth and long-run living standards.
• American Time Use Survey (ATUS): annual survey of how Americans spend their time; the only comprehensive national data source on unpaid work (childcare, eldercare, home production).
• Survey of Occupational Injuries and Illnesses (SOII): annual survey of workplace injuries and illnesses; used for OSHA enforcement targeting.

How It Affects You

If you receive Social Security: Your annual COLA is set each October using the CPI-W for the third quarter of the year, released by BLS in mid-October. In 2025, the COLA was 2.5%, adding roughly $49/month to the average retirement benefit (~$1,976/month). In 2023 it was 8.7% — the largest since 1981 — worth about $140/month for the typical retiree. The COLA can only go up, never down (floor of 0%), but the formula's reliance on CPI-W rather than a seniors-specific index (CPI-E, which BLS produces as an experimental measure) has been a persistent policy debate: seniors spend more on healthcare and housing, which inflate faster than the broad CPI basket.

If you have a TIPS position or I-Bonds: Your returns are directly indexed to BLS CPI-U. A $100,000 TIPS holding adjusts its principal by CPI-U each month; a 3% annual CPI print adds $3,000 to your principal base. I-Bond rates reset every May and November based on the trailing six-month change in CPI-U; the May 2022 I-Bond composite rate hit 9.62%, driving a historic surge in purchases before rates normalized. Watch the March and September CPI releases — they set the six-month I-Bond rate adjustments.

If you're a worker in a unionized industry: About 14 million U.S. union workers have contracts with CPI escalation clauses. A 1-point move in annual CPI-W can translate to thousands of dollars in wage adjustments across multi-year contracts. The UAW, SEIU, and public-sector unions routinely cite BLS CPI data in bargaining — and monitor BLS OEWS wage data to benchmark pay demands against median occupational wages in their metro area.

If you're an employer with federal contracts: BLS Occupational Employment and Wage Statistics (OEWS) set the prevailing wages you must pay on H-1B petitions (Labor Condition Applications require Level I–IV OEWS wages), H-2A agricultural worker programs, and Service Contract Act-covered work. The Davis-Bacon prevailing wage for your county — now applying to IIJA and IRA infrastructure projects — is derived partly from BLS wage surveys. Underpaying relative to BLS wage determinations is a compliance violation that can result in back-pay liability and debarment.

If you work at a federal agency: CPI and ECI data feed directly into the pay comparability calculations underlying General Schedule (GS) pay adjustments. The Federal Employees Pay Comparability Act (FEPCA) requires annual surveys comparing GS salaries to BLS OEWS private-sector wage benchmarks by locality — the gap drives the locality pay percentage for your duty station. If BLS OEWS data shows widening public-private pay gaps in your metro area, your locality pay adjustment (currently ranging from ~16% in rest-of-U.S. to ~34% in San Francisco) may increase in future years.

If you're a journalist or policy analyst: Mark your calendar for the first Friday of each month (jobs report, 8:30 a.m. ET) and the 10th–13th (CPI, 8:30 a.m. ET). Key diagnostic questions: For jobs reports — compare the CES establishment count to the ADP private payroll estimate (released two days prior) and to the CPS household survey count; large divergences signal measurement noise. For CPI — isolate shelter's contribution to core CPI (BLS publishes this breakdown in the release tables); if shelter is still carrying 60–70% of core CPI, the headline overstates underlying inflation. BLS.gov provides the full data API at no cost (JSON, series IDs, no API key for small pulls).

Policy Watch

Data quality and independence are the two fault lines to monitor. The specific risks:

• CPS sample reduction: Any cut to the ~60,000-household CPS sample increases the margin of error on the unemployment rate and makes state-level and demographic breakdowns statistically unreliable. A sample of ~40,000 (a scenario floated in 2025 budget discussions) would roughly double the monthly margin of error from ±0.2 to ±0.3 percentage points — small nationally but material for subgroup analysis (e.g., Black unemployment rate, youth unemployment).
• OEWS coverage gaps: If BLS suspends annual OEWS updates for some metro areas or occupation groups, H-1B prevailing wage determinations and Davis-Bacon calculations revert to older, potentially stale data — systematically underpaying workers in fast-moving labor markets.
• CPI methodology pressure: The Boskin Commission (1996) found CPI overstated inflation by ~1.1 percentage points; subsequent methodological changes reduced this. A future administration could push BLS to adjust CPI weights or switch index formulas, which would directly affect Social Security COLAs and TIPS returns.
• CIPSEA protections: Any legislative attempt to weaken the confidentiality protections for BLS microdata (which contain linked individual-level records across multiple surveys) would likely reduce survey participation rates and degrade data quality across all BLS products.

Independence and DOGE Pressure

BLS statistical independence rests on OMB Statistical Policy Directive No. 1 (SPD 1), which prohibits political appointees from reviewing, altering, or receiving pre-release access to statistical releases. The norm has generally held across administrations — though the Reagan administration attempted to cut the BLS staff in the 1980s, and the Trump administration's first term featured no interference with releases despite broader tensions with economic data agencies.

In 2025, the DOGE initiative proposed significant budget cuts to BLS, including reductions to the Current Population Survey sample size (which would widen the margin of error on the unemployment rate), elimination of some regional data products, and a reduction in OEWS coverage. Statistical methodologists warned that CPS sample reductions would make the unemployment rate statistically unreliable for many demographic subgroups and state-level comparisons. Several statistical agency leaders publicly described DOGE data-access requests — particularly for BLS microdata files linking individual respondents across surveys — as potential violations of CIPSEA's confidentiality protections.

Recent Developments

• 2026 — BLS proposed FY2026 budget reflects flat funding following 2025 rescissions; CPS sample maintained but some state-level occupation data products suspended.
• 2025 — DOGE proposed workforce reductions at BLS; SPD 1 cited as barrier to DOGE access to survey microdata; no changes to release calendar or methodology were implemented.
• 2024 — Annual benchmark revision (March 2024) revised FY2023 nonfarm payrolls down by 818,000 — the largest downward revision since 2009; sparked debate over real-time reliability of jobs data.
• 2023 — BLS published updated CPI shelter methodology documentation responding to widespread criticism of OER lag effects on measured inflation.
• 2022 — CPI reached 9.1% (June 2022), the highest since 1981; BLS CPI releases drove historic Federal Reserve rate-hiking cycle; "supercore" inflation concept emerged from BLS data breakdowns.