America’s economic policymakers are flying blind with outdated instruments. While the Federal Reserve makes important decisions based partially on inflation data, the metrics utilized are plagued by methodological flaws that distort our understanding of the true cost of living. Recent events have made these problems impossible to ignore: shelter inflation remained stubbornly elevated through 2024 even as real-time rental data showed cooling markets, and a government shutdown in October 2025 left us without October CPI data altogether, exposing the fragility of our measurement system. It’s time for a serious conversation about reforming these critical economic indicators.

The Housing Lag Problem: A 17-Month Information Blackout

Perhaps the most glaring issue is how these measures handle housing costs, which represent roughly one-third of the CPI basket and 17% of PCE. Both indices rely heavily on “owner’s equivalent rent” (OER)—essentially asking homeowners what they think they could rent their homes for. This approach introduces a massive lag between real-world housing market changes and what appears in official inflation statistics.

Recent research from the Federal Reserve Bank of Minneapolis reveals the stunning scale of this problem: CPI shelter inflation is best predicted using housing price data from 17 months prior, while PCE housing inflation lags by 19 months. That’s not a typo—our most important inflation measure is operating with information that’s a year and a half out of date.

The real-world consequences have been dramatic. Market rents, measured by indices like Zillow, peaked in early 2022 and returned to pre-pandemic levels by August 2023. Yet CPI shelter inflation peaked 12 months after market rents and was projected to remain elevated into 2025. This meant that through much of 2023 and 2024, the Federal Reserve was raising interest rates to combat housing inflation that had already cooled in the actual market—potentially overtightening monetary policy and risking unnecessary economic pain.

The methodology compounds the problem. The Bureau of Labor Statistics surveys rental units on a staggered six-month schedule, then takes the sixth root of these changes to estimate monthly rates. As the Minneapolis Fed notes, even if monetary policy affected market rents instantaneously, the rents measured in the CPI would only gradually reflect these market conditions. When you’re trying to land a $27 trillion economy, flying with instruments showing conditions from 17 months ago is a recipe for disaster.

Substitution Bias and the Basket Problem

The famous 1996 Boskin Commission estimated that the CPI overstated inflation by about 1.1 percentage points annually, with substitution bias accounting for 0.4 percentage points of that total. The Bureau of Labor Statistics responded by introducing a geometric means formula in 1999 to address “lower-level” substitution (switching between brands of the same product), and created the Chained CPI (C-CPI-U) to capture “upper-level” substitution (switching between product categories).

These reforms helped—current estimates suggest the remaining bias is around 0.5 percentage points per year—but significant problems remain. The CPI’s consumption basket is updated only every two years, meaning rapid shifts in consumption patterns aren’t captured quickly enough. The post-pandemic surge in remote work technology spending, the explosion of food delivery services, and changing transportation patterns all took years to be properly reflected in the official basket.

More fundamentally, there’s a troubling disconnect: the Federal Reserve targets PCE inflation for policy purposes, yet Social Security adjustments and most private contracts use CPI. These measures consistently diverge by about 0.3-0.46 percentage points, with CPI running higher. This creates public confusion and the perception that policymakers are moving goalposts. Americans watching their Social Security checks get CPI-based adjustments while the Fed celebrates hitting its PCE target can be forgiven for wondering if they’re being shortchanged.

Quality Adjustments: A Black Box That Cuts Both Ways

Both indices attempt to adjust for quality improvements through “hedonic” adjustments, but these remain controversial and opaque. The Bureau of Labor Statistics reports that its quality adjustments reduced the new car price index by 80% compared to what it would have been from 1967 to 1994—a staggering figure that raises questions about whether we’re measuring inflation or progress.

The debate cuts both ways. The 1996 Boskin Commission argued that insufficient quality adjustment created a 0.6 percentage point upward bias. But critics counter that modern adjustment procedures may create downward bias by treating all price increases accompanying new features as pure quality improvements rather than inflation. When a laptop costs $1,200 instead of $1,000 but has a faster processor, statisticians might calculate zero inflation. While theoretically sound, this makes official statistics feel disconnected from the lived experience of families paying 20% more for a “necessity.”

The methodology is particularly problematic for healthcare and education, where quality improvements are difficult to measure and consumers often lack meaningful choice. Did your health insurance premium increase by 8% because of inflation, or because insurers now cover telemedicine? The answer matters enormously for policy, yet the adjustments are rarely transparent to the public or even to most policymakers. As the Boskin Commission acknowledged, these calculations involve subjective judgment calls that can swing inflation estimates by half a percentage point or more annually.

Core vs. Headline: Missing the Forest for the Trees

The Federal Reserve’s focus on “core” inflation (excluding food and energy) made sense when these prices were volatile but mean-reverting. However, this framework breaks down during structural shifts. Through 2024, core CPI fluctuated between 3.2% and 3.4% while headline inflation ranged from 2.7% to 3.3%, creating public confusion about whether inflation was truly under control. By December 2024, headline CPI stood at 2.9% while core remained at 3.2%—but families don’t eat “core” inflation.

The White House Council of Economic Advisers has acknowledged this challenge, noting that housing inflation remained the dominant driver of core CPI persistence, accounting for 2.3 percentage points of the 3.3% core inflation rate through mid-2024. If shelter inflation were at pre-pandemic levels, core CPI would have been near target—but that’s cold comfort to families facing actual elevated housing costs.

Moreover, the assumption that food and energy prices are merely “volatile noise” looks increasingly questionable. When food price spikes stem from climate change-driven droughts rather than temporary weather, or when energy transitions create persistent price shifts, excluding these categories may blind policymakers to real structural changes in the cost of living.

The Deeper Problem: Should Policy Target Economic Purity or Consumer Pain?

This raises a fundamental question: should the Federal Reserve optimize its policy for a “purer” measure of underlying inflation trends, or for the actual pain consumers experience at the grocery store and gas pump?

The Fed’s official inflation target is actually 2% for headline PCE—the measure that includes food and energy. Yet in practice, Fed communications obsessively focus on core measures. This creates a troubling disconnect. As former St. Louis Fed President James Bullard wrote, the Fed should target headline “because it makes sense to focus on the prices that U.S. households actually have to pay.” Less-affluent households, which spend disproportionate shares of their budgets on food and energy, experience inflation very differently than the core measure suggests.

The traditional justification for focusing on core inflation has two related but distinct arguments. First, food and energy prices are volatile—they swing wildly month to month, making trends hard to discern. Second, and more fundamentally, these price changes stem from exogenous supply shocks that monetary policy cannot address. When a drought reduces the wheat harvest or OPEC cuts oil production, raising interest rates won’t make it rain or pump more oil. These are relative price changes (one good got more expensive) rather than true inflation (all prices rising together). The theory says if these shocks are temporary and don’t trigger “second-round effects” into other prices, the Fed should “look through” them rather than slowing the economy unnecessarily.

But this elegant theory rests on assumptions that are looking increasingly shaky in the 2020s:

The shocks aren’t always temporary anymore. Climate change is causing persistent food price pressures, not one-off weather events. Energy transitions may drive sustained price changes. When “shocks” last for years, they’re no longer noise to filter out—they’re part of the actual inflation signal.

Not all food and energy price changes are supply shocks. Recent research decomposing food inflation into supply and demand components shows that demand-driven food inflation responds to monetary policy—but the Fed can’t distinguish this in real-time when reflexively focusing on core. By excluding all food price changes, the Fed may be ignoring price pressures it actually can and should address.

Second-round effects are real and persistent. The European Central Bank’s recent analysis shows that when food and energy prices spike, they DO feed into other prices through wage indexation and shifted inflation expectations, especially for highly visible prices that shape consumer psychology. The Boston Fed’s historical analysis confirms this: when energy prices rise persistently, headline and core diverge for years rather than converging as theory predicts.

History provides a stark warning. In the 1970s, the Fed focused on what it thought was “underlying” inflation while dismissing persistent energy price increases as temporary supply shocks. The result was runaway inflation that took a brutal recession to tame. A similar dynamic may have occurred from 2003-2007, when accommodative policy looked reasonable based on benign core inflation even as headline surged—potentially overheating the economy.

Research on emerging economies illuminates what’s at stake. In countries where food represents a larger budget share, food price shocks have demonstrable long-run effects and cannot be “looked through”—they must be addressed. While the US is wealthier, lower-income American households still spend 30-40% of their budgets on food and energy. For them, the distinction between core and headline isn’t academic—it’s the difference between eating and heating their homes.

There’s something deeply troubling about a central bank telling families that inflation is under control because core measures look good, while those same families are paying 36.8% more for eggs and watching their grocery bills surge. This isn’t just a technical disagreement among economists—it’s a question of whose economic reality matters. The Fed’s mandate is to promote maximum employment and stable prices for actual Americans, not to hit targets for economically convenient statistical constructs that exclude the items Americans consume most frequently.

The PCE vs. CPI Confusion

While the Federal Reserve targets PCE inflation for policy purposes, most Americans are more familiar with the CPI, which is used for Social Security adjustments and many private contracts. The two measures can diverge significantly—PCE typically runs about 0.3 percentage points lower than CPI—creating public confusion and the perception that policymakers are moving goalposts.

This divergence stems from different methodologies, weights, and coverage, but the average citizen can be forgiven for wondering why we have two competing measures of something as fundamental as inflation.

A Path Forward: Specific Reforms We Need Now

Reform is both necessary and achievable. Here’s what needs to happen:

Modernize housing measurement immediately. The 17-month lag in housing data is unacceptable in the age of real-time information. The Bureau of Labor Statistics and Bureau of Economic Analysis should integrate real-time rental data from private sources like Zillow, Apartment List, and RealPage, with appropriate quality controls and adjustments. The Cleveland Fed has already demonstrated that such integration is feasible through its inflation nowcasting models. The technology exists—we just need the political will to implement it.

Update consumption baskets quarterly, not biannually. Modern data collection techniques make quarterly updates not just feasible but necessary. The pandemic proved that consumption patterns can shift radically in months, not years. Our inflation measures need to keep pace.

Make quality adjustments transparent and reviewable. Every quality adjustment above a certain threshold—say, 2% of an item’s price—should be publicly documented with clear methodology. An independent board should review major adjustments annually. The public deserves to understand why their grocery bill is up 15% but “quality-adjusted” food inflation is supposedly lower.

Harmonize CPI and PCE, or explain the divergence clearly. The current system, where Social Security uses CPI but the Fed targets PCE, creates unnecessary confusion. Either align the measures, or have the Fed and Social Security Administration jointly publish an annual report explaining why the measures differ and what that means for beneficiaries. The Chained CPI (C-CPI-U) already exists and more accurately accounts for substitution—serious proposals to use it for Social Security COLAs should be considered, as it would close about one-sixth of the program’s funding gap while providing a more accurate cost-of-living measure.

Develop complementary inflation measures for different income groups. A “necessities index” focusing on non-discretionary spending—housing, food, healthcare, energy, education—would help policymakers understand cost-of-living pressures on lower-income families who spend a larger share of income on these essentials. The current indices treat all consumers as having identical spending patterns, which is fiction.

Build resilience into the measurement system. The October 2025 government shutdown left us without October CPI data, exposing dangerous fragility in our statistical infrastructure. The BLS needs emergency protocols and the ability to reconstruct data from alternative sources when primary collection fails. Too many critical economic decisions depend on these numbers for us to accept such gaps.

Invest in the statistical agencies. None of this happens without adequate funding. The BLS and BEA have seen their budgets squeezed for years while being asked to produce ever more sophisticated measures. Congress needs to provide the resources necessary for 21st-century inflation measurement, including hiring more economists, upgrading data systems, and expanding survey samples.

The Stakes Are Higher Than Ever

These aren’t merely technical quibbles for econometricians. Inflation measurement affects monetary policy decisions that impact millions of jobs, determines Social Security payments for tens of millions of seniors, influences wage negotiations across the economy, and shapes public trust in government institutions. When our measures lag reality by 17 months or fail to capture true cost-of-living changes, we risk catastrophic policy mistakes.

The recent period has laid bare the consequences of our flawed measures. The Federal Reserve may have kept rates higher for longer than necessary because shelter inflation statistics showed heat that had already dissipated in actual markets. Families have watched their purchasing power erode while being told inflation is moderating—because the quality adjustments say so. The October 2025 government shutdown revealed that a single month’s lapse in appropriations can leave us flying completely blind.

The Bureau of Labor Statistics and Bureau of Economic Analysis do heroic work with the tools and mandates they have, but those tools are antiquated for our modern, rapidly-changing economy. We have the technology to measure inflation more accurately, more frequently, and with more granularity than ever before. What we lack is the political commitment to fund and implement these improvements.

It’s time for Congress to treat statistical infrastructure with the seriousness it deserves—funding adequate modernization, mandating quarterly basket updates, requiring transparency in quality adjustments, and building redundancy against measurement failures. For economists and policymakers, it’s time to engage honestly with the limitations of our current measures and stop pretending that decades-old methodologies are fit for purpose in 2025.

America’s families deserve inflation statistics that reflect their reality, not a lagging, smoothed, quality-adjusted approximation that serves the convenience of policymakers more than the needs of the people. The cost of continuing with broken measures far exceeds the investment needed to fix them.

And fixing measurement is just the first step. Even with better data, we’ll still need to wrestle with equally important questions about the Fed’s targeting framework itself—whether 2% inflation should be a point target or averaged over time, and over what horizon. But we can’t have that conversation productively until we’re confident we’re measuring inflation accurately in the first place. Let’s start by getting the measurement right.