The Aggregate

There is a business-cycle observation, dating to Wesley Mitchell and the earliest research of the National Bureau of Economic Research, that recessions originate in a small number of interest-rate-sensitive sectors. The modern version of the claim — formalized by economist Wallace Duncan and circulated widely among portfolio managers and market analysts — is that roughly twenty percent of the economy drives one hundred percent of recessions.

The thesis identifies three sectors — durable goods consumption, residential investment, and business equipment investment — as the components that generate nearly all cyclical fluctuation in GDP. Track them, and you track the cycle. Ignore them in favor of aggregate indicators — total GDP, total consumption, "the consumer" — and you are following, not leading. The non-cyclical eighty percent of the economy almost never contracts. Conventional analysis, by focusing on the stable majority, is structurally late to every inflection point.

The diagnosis is widely accepted among practitioners. It also carries an implicit assumption about the Federal Reserve: that the institution responsible for monetary policy does not track these sectors with adequate precision, or, if it does, that it cannot act on what it sees. Thousands of investment professionals now build their cycle-timing around this assumption.

These are testable claims — not against economic theory, but against the documentary record of the institution itself. And the test has a second layer the practitioners' framework does not anticipate. The Federal Reserve does not merely track the cyclical sectors. It monitors the financial conditions that drive them — credit spreads, lending standards, liquidity indicators — through instruments like the Senior Loan Officer Opinion Survey that provide upstream signal quarters before the sectoral GDP data turns. If the committee possessed both the sectoral decomposition and the financial conditions channel that leads it, and still produced the aggregate narratives the practitioners criticize, then the failure is not one of detection. It is something the framework has not diagnosed.

We went into the archives.

The Machinery Behind the Index

The Federal Reserve's internal forecasting documents have employed a standardized template — "Net Changes in Real Gross Domestic Product and Related Items" — that explicitly isolates the three sectors the cyclical-GDP thesis identifies as the keys to the business cycle. Since at least 1980, every Greenbook and Tealbook presented to the FOMC has broken out "Personal consumption expenditures: Durables," "Business fixed investment: Producers' durable equipment," and "Residential structures" as distinct line items with their own growth projections, contribution calculations, and risk assessments.

But the Fed's apparatus does not merely track these sectors as share-of-GDP indicators. It models the transmission channels that make them cyclical. The FRB/US model — the Fed's workhorse macroeconomic model — contains full behavioral equations for each interest-rate-sensitive sector. Residential investment in FRB/US responds simultaneously to mortgage rates, housing demand shocks, wealth effects, and credit conditions. Business equipment investment responds to the cost of capital, expected output growth, and financial accelerator mechanisms. The model captures why these sectors are cyclical — they are the channels through which the committee's own policy instrument transmits to the real economy — rather than merely observing that they are.

David Stockton, who served as Director of the Division of Research and Statistics — the Fed's chief forecaster — articulated the framework's rationale to the committee on December 12, 2006:

By 2011, the staff formalized this decomposition into the EDO model — Estimated Dynamic Optimization-based — explicitly designed to capture the "cyclical properties of different categories of durable expenditure." The practitioner framework is a heuristic index that observes sectoral shares. The Fed's apparatus is a full general-equilibrium model that explains sectoral behavior through the interest-rate channel the committee controls.

There is a further layer the outside framework does not reach. The staff tracks a set of financial conditions indicators — the Senior Loan Officer Opinion Survey, credit spreads, bank lending standards, term premia — that lead the sectoral movements outside indices track. The documentary record shows this upstream channel operating with measurable lead times. In March 2000, Peter Fisher, Manager of the System Open Market Account, warned the committee that credit spreads had widened to their 1998 highs and that spread relationships among private credits had "become much less stable." By May 2000, staff reported that junk-rated borrowers had been "shut out of the market" for convertible bonds. Equipment investment data would not collapse for another six months. In June 2005, Governor Susan Bies told the committee that staff were "too sanguine" about the momentum building in the housing sector, drawing on supervisory data about weakened lending standards that the Senior Loan Officer Survey would not formally register until April 2007. Residential investment would not begin its double-digit decline until the following year.

The staff is not watching the cyclical sectors in isolation. It is watching what drives the cyclical sectors. And the documentary record shows the upstream financial conditions channel providing six to thirty months of additional lead time beyond the sectoral GDP data alone. This makes the subsequent institutional failure not merely more damning but categorically different from what the practitioners' heuristic can identify. An analyst who lacks the Fed's internal machinery might reasonably miss a cyclical turning point. An institution that possesses behavioral models for each sector, tracks the financial conditions that lead those sectors by quarters, and has a chief forecaster who articulates the methodology on the record — that institution's failure to follow through requires a different explanation entirely.

The Equipment Collapse

The 2001 recession offers the cleanest test because the downturn was concentrated in a single cyclical component. Business equipment investment — specifically information technology and telecommunications — collapsed while residential investment actually rose. If the FOMC was engaging with both layers of its apparatus — the sectoral decomposition and the upstream financial conditions — committee members should have been discussing equipment-specific deterioration as a leading cyclical signal, and the credit conditions that preceded it. If they were converting the staff's precision into aggregate narrative, they would have been talking about "the economy" or "the consumer" while both layers of leading indicator were already flashing.

The upstream financial conditions channel fired first. In March 2000, Peter Fisher reported to the committee that ten-year credit spreads had widened to 1998 levels and that spread relationships had become "much less stable." By May 2000, the Greenbook reported that junk-rated borrowers had been "shut out" of the convertible bond market by falling share prices and rising volatility, and the Senior Loan Officer Survey showed twenty-five percent of domestic respondents tightening standards on commercial and industrial loans. The credit channel through which speculative-grade technology firms financed their equipment purchases was closing.

But the staff's baseline forecast did not incorporate what the staff's own financial conditions tracking was showing. In August 2000 — three months after the junk bond market shut down — staff reported that business equipment spending "continued to advance at a rapid clip," increasing at a twenty-one percent annual rate. In January 2000, the Greenbook had projected computing equipment outlays would increase fifty to sixty-five percent over the following two years. The financial conditions channel was signaling that the financing base for this growth had collapsed. The baseline forecast projected the growth would continue.

David Stockton — the same analyst who would later articulate the cyclical-turning-point methodology — acknowledged the tension in October 2000 without resolving it:

By June 2001, "nervous" had become "bleak." Staff reported that high-tech orders had plummeted forty-five percent and that communications equipment shipments had been declining for months. The upstream signal — credit spreads, junk bond market closure, SLOOS tightening — had preceded the sectoral collapse by six to eight months. The staff had possessed both layers of warning and followed neither to its conclusion.

Individual committee members engaged with the equipment-specific signal. Vice Chairman Roger Ferguson identified the commercial paper market as "anything but healthy" in January 2001, reading the upstream financial conditions channel directly. Vice Chairman William McDonough warned in April 2001 that business executives were delaying investment decisions and that "the downside risk to our economic forecast is extremely considerable." But the committee consensus reached for the aggregate buffer. Governor Edward Kelley dismissed the technology downturn as "another episode of sectoral rolling recession" that "seldom drags down the entire economy." When consumer spending appeared resilient in late 2001, the committee attributed it to structural consumer strength — though staff private analysis showed the spending jump was driven by zero-rate auto financing packages, a temporary durable goods stimulus, not evidence of aggregate resilience.

The pattern had established itself across both layers: the upstream financial conditions signal detected the credit channel closing, the sectoral data confirmed the investment collapse months later, the staff's baseline forecast projected growth through both warnings, individual members engaged with the signals by name, and the committee consensus absorbed everything into an aggregate narrative that required no policy change. But the 2001 recession was mild. The equipment collapse was concentrated in a single cyclical component while others held. The test of the pattern — and both layers of it — was whether it would repeat when the leading sector was larger, the financial conditions deterioration more systemic, and the consequences of late recognition catastrophic.

The Housing Subtraction

If the 2001 episode established the pattern, the 2006-2007 housing decline revealed what happens when both layers of the Fed's apparatus fire with maximum urgency and the committee's consensus process converts both into containment narrative. The upstream financial conditions channel activated earlier, the sectoral deterioration was larger, and the gap between what the staff's instruments were showing and what the committee told the public was wider than in any episode the archive records.

The upstream layer fired first — and from inside the committee itself. In June 2005, Governor Susan Bies challenged the staff directly:

At the same meeting, Governor Mark Olson identified the mechanism that made the housing exposure systemic rather than sectoral — the compounding of individually manageable risks into something no single risk measure could capture:

Bies's warning came from supervisory data — her role overseeing bank supervision gave her access to information about weakened proof-of-income and appraisal standards that the Senior Loan Officer Survey would not formally register for nearly two years. Olson's "layering" concept identified the transmission channel the FRB/US model was designed to capture: not any single risk, but the compounding of low documentation, high loan-to-value ratios, and variable rates into a vulnerability that could not be assessed by looking at any component in isolation. These were upstream financial conditions warnings, delivered thirty months before the recession began, by members who were reading the credit channel that no outside sectoral index can replicate.

The sectoral layer confirmed the upstream signal on schedule. By June 2006, the Greenbook documented a downward revision of almost six percentage points in real residential investment, projecting a seven percent annual rate decline. By August, that projection had worsened to negative fourteen percent. The staff's Greenbook told the committee that residential investment was subtracting 1.2 percentage points from GDP growth. The financial conditions section of the same Greenbook was showing mortgage credit tightening and novel financing risks. Both layers of the Fed's apparatus — the upstream credit conditions and the downstream sectoral data — were aligned.

The committee's response was to quarantine both signals. Michael Moskow, President of the Federal Reserve Bank of Chicago, provided the formulation:

By July 2007, the gap between the committee's public posture and the staff's own survey instrument had become a documentary contradiction. The committee's public statement on July 18, 2007, asserted that "conditions in financial markets have remained generally supportive of economic expansion." At the same moment, the staff's Senior Loan Officer Survey showed fifty-three percent of domestic respondents had tightened subprime mortgage lending standards, and the April 2007 survey had already documented tightening of commercial real estate and commercial and industrial lending — non-housing sectors. The committee's own financial conditions survey was contradicting the committee's own public financial conditions assessment.

By October 2007, the contradiction deepened further. Banks told the Senior Loan Officer Survey that they were tightening C&I lending standards because of a "less favorable economic outlook" — not because of housing, not because of subprime, but because they believed the broader economy was weakening. The staff's upstream instrument was reporting that the containment narrative had already failed. At the same meeting cycle, President Richard Fisher asserted there were "as yet no appreciable, let alone debilitating, signs of spillover into the rest of the economy." The staff possessed two layers of evidence that the spillover had occurred. The committee's public narrative denied both.

The staff's alternative simulations — the "Greater Housing Correction," the "Credit Crunch" scenario projecting one percent GDP growth, the "Bank Capital Crunch" model — proved more accurate than the optimistic baseline. The committee retained the baseline. The economy entered the worst recession since the 1930s. But the question the documentary record now forces is not merely whether the committee aggregated away the sectoral signal — it is whether the aggregation followed a pattern, and whether that pattern was directional.

The Asymmetry

If the committee aggregated all information symmetrically — treating both strength and weakness in broad terms regardless of direction — the finding would be institutional inertia. But the documentary record reveals something more specific. When interest-rate-sensitive sectors were strong, committee members cited them by name. When the same sectors were weak, the committee shifted to aggregate formulations. And when the upstream financial conditions channel was signaling distress, the committee found a specific rhetorical device to dismiss it: majority-weighted aggregation.

The device operated identically in both episodes. In November 2000, as the credit conditions that sustained the technology investment boom were collapsing, a staff member named the principle explicitly during committee deliberations. Mr. Oliner argued that since the investment-grade borrowers represented roughly eighty percent of the market, the committee should put "more weight on what is happening in the investment grade part of the market where conditions don't look nearly as adverse as in the speculative grade part." The logic was precise: the twenty percent of borrowers experiencing credit distress — the speculative-grade, IPO-dependent technology firms that had driven the investment boom — could be excluded from the assessment because they were a minority of the total.

Six years later, Moskow applied the same structure to the housing episode: the five percent of the economy experiencing contraction — residential investment — could be excluded because ninety-five percent remained "on good footing." The device was identical. Identify the leading signal. Note that it represents a minority — of borrowers, of GDP, of the economy. Weight the majority. Conclude the aggregate is sound. In both cases, the minority being excluded was the component that had driven the preceding expansion and whose contraction would determine the cycle. The staff's own analysis had identified this. The committee's majority-weighted aggregation dismissed it.

The purest expression of the asymmetry was the "handoff" theory. Jeffrey Lacker, President of the Federal Reserve Bank of Richmond, argued from late 2005 through 2006 that the economy would execute a smooth transition from housing-led growth to business-investment-led growth:

The handoff theory named one cyclical sector's strength as the instrument of offset while absorbing another cyclical sector's weakness into the aggregate. The committee never applied this logic in reverse. No member argued during the housing boom that residential investment strength would offset a potential decline in business equipment. The asymmetry was systematic: specific attribution for good news, aggregate absorption for bad news, and the net conclusion always favoring the narrative that required no policy change.

This is the familiar pattern — the "the consumer is fine" narrative that persists through early recession stages — but produced not by outside commentators. It was produced by the institution itself, by members who had just received a Greenbook showing the cyclical sectors in decline and a Senior Loan Officer Survey showing the credit conditions that drive those sectors tightening across non-housing categories. The committee was not repeating what financial media said. It was originating the narrative. And the aggregation dismissed not one but two layers of the Fed's own apparatus — the sectoral data and the upstream financial conditions that led it.

If the asymmetry was this consistent — strong sectors named, weak sectors absorbed, upstream financial conditions reframed as temporary friction — the question becomes whether anyone inside the committee recognized the pattern and resisted it. If the aggregation was total, the finding is institutional. If individuals challenged the containment directly, reading both layers of the Fed's own instruments, and were overridden, the finding is sharper.

The Dissenters Who Decomposed

The aggregation was not unanimous. A persistent minority of committee members did exactly what the staff's analytical framework prescribed: they read the upstream financial conditions, they tracked the sectoral decomposition, and they identified the cyclical sectors as leading signals rather than contained adjustments. The documentary record reveals a hierarchy among these dissenters — those who read the upstream credit channel were more prescient than those who read the sectoral data, and both were more prescient than the consensus that aggregated everything away.

The earliest warnings came from the upstream layer. Bies and Olson in June 2005, reading supervisory data and lending standards that the SLOOS would not formally register for two years. Peter Fisher in March 2000, reading credit spread instability that preceded the equipment investment collapse by eight months. Vice Chairman Roger Ferguson in January 2001, identifying the commercial paper market as "anything but healthy" while the committee consensus still treated the technology downturn as a contained sectoral adjustment. These voices were reading the financial conditions channel that outside sectoral indices do not capture — and they were right earlier than anyone tracking the sectors alone.

The sectoral dissenters followed, engaging with the GDP decomposition the staff produced. Anthony Santomero, President of the Federal Reserve Bank of Philadelphia, provided the first quantified challenge in January 2006:

By December 2006, Sandra Pianalto, President of the Federal Reserve Bank of Cleveland, positioned herself directly against the handoff consensus:

Chairman Bernanke's response, at the same meeting — the same meeting where Stockton told the committee that cyclical turning points appear first in durable goods and construction — was the aggregate formulation: "most people around the table accepted the general contour of the Greenbook forecast — that is, moderate growth... So far there is little evidence of spillover into consumption in particular."

Through 2007, the upstream and sectoral warnings converged. Janet Yellen flagged wealth-effect spillovers in October 2006, worried that house prices would actually fall while the Greenbook assumed they would be flat, and by August 2007 identified the systemic mechanism:

Eric Rosengren delivered the single most diagnostic observation of the period — one that identified not just the risk but the reason the institution's own instruments were being ignored:

Rosengren was identifying a model failure — the disconnect between what the staff's econometric framework projected and what the financial conditions channel was actually transmitting. The "benign outlook of the forecasters" was the mean-reversion baseline that overwrote both layers of warning. The "angst" was the upstream signal that the staff's own SLOOS data confirmed but the baseline could not process. By December 2007, Rosengren articulated the structural point: "disruptions in financial flows have the potential to result in significantly more weakness than would result from an econometric model." The staff's own May 2000 analysis had said the same thing seven years earlier: formal models "routinely fail to capture the sorts of 'nonlinearities' that come into play once the economy starts to decelerate appreciably," which is why the Fed has previously "ended up skidding into the ditch."

And then there was Lacker. The architect of the handoff theory watched its foundation erode. By October 2007, he admitted he was "at least as pessimistic as the Greenbook about prospects for housing bottoming out." By January 2008, he reported firms in his district delaying capital spending — the very investment that was supposed to provide the offset. On March 18, 2008:

The man who had named business capital spending as the bright spot that would offset residential investment's decline was now declaring a recession that the handoff was supposed to prevent. The cyclical components had declined together, as they do in every serious downturn. The dissenters who had argued that residential investment weakness was a leading cyclical signal, not a contained sectoral adjustment, had been right. The upstream voices — Bies, Olson, Rosengren — who had read the financial conditions channel had been right earlier. The consensus that had aggregated both layers of warning into moderate growth, underlying strength, and the ninety-five percent of the economy outside of housing had been wrong.

The question this leaves is whether the aggregation is correctable. The evidence suggests it is not — or at least, not without restructuring the consensus process itself. The staff built both layers of the analytical instrument. Stockton articulated the sectoral methodology. The SLOOS tracked the upstream credit conditions. Individual members read both layers correctly. And still the committee consensus produced aggregate narratives that obscured what every layer of the institution's own analysis was showing. The committee did not merely ignore the warning. It reframed the warning's meaning. The staff's October 2007 finding that banks were tightening because of a "less favorable economic outlook" — a fundamental economic judgment — was transformed in public communications into "difficulties in valuing complex or illiquid financial products" — a technical market-friction problem. The upstream signal was not suppressed. It was translated from an economic verdict into a financial plumbing issue. Poole identified the institutional constraint in August 2006: "we don't want the market to believe that policy is really being driven by one sector of the economy." This is not an analytical failure. It is a communication doctrine — a commitment to projecting broad-based assessment that structurally resists the specificity both layers of the staff's own framework require. (The Paradox documented the same structural lock-in from the opposite direction: when a single inflation component ran inconveniently high, the committee found technically valid reasons to exclude it from the measure that guided policy. The mechanism operates in both directions — excluding what is inconveniently weak and excluding what is inconveniently strong — and in both cases the exclusion serves the narrative that requires no change.)

The Coarse Material