State-of-the-Art Macroeconomics (2024–2026)

A foundational, encyclopedic survey of where research-frontier macroeconomics stood circa 2024–2026: what was settled, what was contested, and where the field was visibly moving. Written for a serious self-learner who is reading a parallel macro-history report (see Cross-references at the end).

Executive summary

Macroeconomics in the mid-2020s is in a peculiar moment. The post-COVID inflation surge of 2021–2023, the most synchronized global tightening cycle in a generation, and the rapid expansion of central bank balance sheets forced the field to revisit assumptions that had hardened during the long, low-inflation decade after the Global Financial Crisis (GFC).

The result is not a paradigm shift in the Kuhnian sense. The basic skeleton — a New Keynesian (NK) model where central banks set interest rates to manage an inflation–activity trade-off — remains the workhorse. But four major retoolings are underway. (1) The representative agent has been substantively displaced by heterogeneous-agent (HANK) models that take seriously who holds wealth and who lives "hand-to-mouth." (2) The Phillips curve has been rehabilitated from a flat, dormant relation into a nonlinear, state-dependent one. (3) The fiscal–monetary frontier has reopened: the Fiscal Theory of the Price Level (FTPL), once a fringe contender, is now a respectable, if contested, framework for explaining the 2021–2023 inflation. (4) Central bank balance sheets are no longer a temporary crisis tool; "ample reserves" frameworks, standing repo facilities, and central-bank-financed losses have become a permanent operating environment, raising questions about quasi-fiscal independence.

Settled or near-settled: inflation expectations matter enormously (and were better anchored than pessimists feared); the zero lower bound (ZLB) constraint is no longer the modal scenario assumed in policy design; productivity slowdowns in the 2010s were real, broad-based, and only partly explained by mismeasurement; the trilemma is more like a "dilemma" in a world of correlated capital flows; fiscal multipliers are positive and meaningfully above 0.5 in slack economies, though the upper bound is contested.

Contested or actively moving: the slope and nonlinearity of the Phillips curve; whether r-star (the equilibrium real interest rate) is genuinely higher than its pre-COVID lows or just temporarily elevated; the share of post-2021 inflation attributable to supply, demand, fiscal, or markup channels; how much of the disinflation of 2023–2024 was "immaculate"; how to design monetary frameworks for an asymmetric world where overshoot risk now feels symmetric with undershoot risk; and the proper role of climate, demographics, and AI in long-run growth and r-star modeling.

This report walks through each of these debates with citations to primary papers and central bank documents, presents the algebra where needed, and flags where deeper mathematical tooling (stochastic calculus, dynamic programming, functional analysis) is required to follow the frontier literature.

Research brief

• Topic. The frontier of macroeconomic research and policy, with emphasis on developments since 2021.
• Audience. A capable self-learner without formal econ training, reading a foundational history of the field in parallel.
• Decision supported. Building durable foundational knowledge that will serve as base for downstream reading in trading-relevant macro, micro-macro intermingling, and ongoing journal/working-paper consumption.
• Recency. Heavy weight 2021–2026; moderate weight 2016–2020; classical citations only when load-bearing.
• Math. Algebra in the body; named-but-not-derived for stochastic calculus, dynamic programming, projection methods, sequence-space Jacobians, Kalman filtering.
• Source policy. Fed/ECB/BIS/IMF working papers and speeches; NBER WPs; AER, AEJ:Macro, JEP, JPE, QJE, Econometrica; Brookings Papers; Jackson Hole; Liberty Street Economics; selected serious blogs (Cochrane, Krugman, Blanchard, DeLong, Mason, Politano/Apricitas, Amarnath/Employ America).
• Out of scope. Detailed history of macroeconomic thought (covered in the parallel history report); finance-theoretic asset-pricing detail (some overlap noted, but the trading-economics report goes deeper); microeconomics frontier (separate SOTA micro report).

Key findings

1. The DSGE workhorse is alive, but heterogeneity is now first-class. The Smets–Wouters New Keynesian DSGE remains the central bank reference model, but Kaplan–Moll–Violante's HANK and Auclert–Bardóczy–Rognlie–Straub's sequence-space Jacobian methods now provide the credible alternative for serious policy analysis.

2. The Phillips curve is not dead — it's nonlinear. The Hazell–Herreño–Nakamura–Steinsson (HHNS) result that the curve was already very flat in the 1980s remains influential, but Benigno–Eggertsson (2023, 2024), Cerrato–Gitti, and ECB Phillips-curve work have re-established that the slope steepens sharply when the vacancy/unemployment (V/U) ratio is high.

3. Post-pandemic inflation was supply-led at first, demand-supported later. Bernanke–Blanchard's 11-economy decomposition, Shapiro's SF Fed series, and Goolsbee's "lessons from the supply side" all support a sequenced reading: commodity and sectoral shocks first, then a tight-labor and well-anchored-expectations regime that allowed disinflation without the deep recession pessimists predicted.

4. The fiscal theory of the price level is no longer a fringe view. Cochrane's 2023 book and Bianchi–Faccini–Melosi (QJE 2023) push a partially fiscal explanation for the 2021–2022 surge. Reis, Bernanke–Blanchard, and others resist, but the FTPL is now part of the standard toolkit.

5. The Fed's 2025 framework review junked Flexible Average Inflation Targeting (FAIT). Powell's August 2025 Jackson Hole speech and the FOMC's revised Statement on Longer-Run Goals restored Flexible Inflation Targeting (FIT), removed the "shortfalls" employment language, and de-emphasized the ZLB as a defining feature.

6. The ECB confirmed its 2% symmetric target in the 2025 strategy assessment. A more volatile world is now baked into the strategy; periodic reviews continue (next 2030).

7. The Bank of Japan exited Yield Curve Control in March 2024. Governor Ueda ended seven years of YCC and 17 years of negative rates, marking the close of the largest experiment in unconventional monetary policy.

8. Central bank balance sheets are durably large. The Fed ended QT in December 2025; the ECB shrunk but did not zero out; the BoE's APF is producing tens of billions in losses indemnified by the UK Treasury. "Ample reserves" is now a permanent regime.

9. r-star estimates have not converged. The Holston–Laubach–Williams (HLW) update places real r-star around 0.7–0.9% (mid-2025); Williams himself argues the "low-r-star regime endures." But Richmond Fed staff and survey-based measures show r-star above 2%. The disagreement is real and has direct implications for whether current rates are restrictive.

10. The Beveridge-curve debate had a winner. The Blanchard–Domash–Summers (BDS) pessimists predicted a hard landing; Figura–Waller (FRB) and Benigno–Eggertsson predicted a soft landing along the vertical Beveridge curve. Through 2024–2025, the soft-landing camp won, with vacancies falling and unemployment rising only modestly.

11. The "savings glut of the rich" and demographic forces shape r-star together. Mian–Straub–Sufi tie low rates to top-end inequality; Auclert–Malmberg–Martenet–Rognlie tie them to demographic aging; Goodhart–Pradhan argue the demographic reversal is now flipping the sign.

12. Climate macro is becoming standard. DICE-2023, the NGFS scenarios (Phase V, November 2024), and the Golosov–Hassler–Krusell–Tsyvinski sufficient-statistic carbon-tax formula have moved climate from a specialist sub-field to a required component of long-horizon central bank scenario work.

13. The "greedflation" debate produced more heat than empirical consensus. Weber–Wasner's seller's-inflation hypothesis and Konczal–Lusiani's markup decomposition documented real margin expansion in 2021; subsequent work showed margins normalizing, and the causal direction (markups → inflation vs inflation → markups in a coordinated game) remains contested.

14. Crypto and stablecoins have moved into macro relevance. Stablecoin assets exceeded $250bn by early 2025; Tether is among the largest single Treasury holders; the US GENIUS Act (July 2025) imposed 100%-Treasury-backed reserves on payment stablecoins. CBDCs remain mostly research-stage in advanced economies but are operational in several emerging markets.

15. Central bank independence is under measurable strain. The Trump administration's pressure campaign on Powell in 2025–2026 is the most public episode in decades. Prediction-market evidence and yield-curve responses suggest markets price the threat to independence as raising long-term yields even as it lowers short-rate expectations.

Detailed analysis

1. The current frameworks: DSGE, HANK, TANK, FTPL, RBC

1.1 New Keynesian DSGE — the workhorse

Dynamic Stochastic General Equilibrium (DSGE) models have been the dominant quantitative framework in macroeconomics since the 1990s. Their lineage runs from the Real Business Cycle (RBC) models of Kydland–Prescott (1982) and Long–Plosser (1983), to the New Keynesian (NK) synthesis with nominal rigidities (Clarida–Galí–Gertler 1999), to the medium-scale estimated NK models of Christiano–Eichenbaum–Evans (2005) and Smets–Wouters (2003, 2007).

The Smets–Wouters (2007) AER paper became the de facto reference. Its model includes:

• Sticky prices (Calvo-style staggered price-setting)
• Sticky wages
• Habit formation in consumption
• Investment adjustment costs
• Variable capital utilization
• A Taylor-type monetary rule
• Seven structural shocks (technology, risk premium, investment-specific, government spending, monetary policy, price markup, wage markup)

The model is estimated by Bayesian methods on seven US time series (output, consumption, investment, prices, real wages, employment, the nominal rate). It tracks and forecasts US macro data competitively with Bayesian VARs (Smets & Wouters AER 2007).

The DSGE workhorse has been heavily criticized after the GFC for: (a) representative-agent assumptions that hide distributional dynamics; (b) reliance on Calvo pricing as an unrealistic micro-foundation; (c) post-hoc shock proliferation; (d) failure to predict the GFC; and (e) the "forward guidance puzzle" (the model predicts implausibly large effects of credibly committing to low future rates) (Bruegel blog: The DSGE Model Quarrel (Again)).

Defenders (Christiano, Eichenbaum, Trabandt 2018 JEP; Linde 2018) argue DSGE is the only tool that imposes general-equilibrium discipline and lets central banks attribute macro outcomes to identifiable shocks. The critique that "DSGE is the only quantitative tool" is straw-manning — alternatives include structural VARs (sVARs) and traditional semi-structural simultaneous-equation models (SEMs), each with comparative advantages.

State of play (2026). Smets–Wouters remains the reference benchmark at the Fed (FRB/US is a sister semi-structural model) and ECB (NAWM-II). But policy committees now consult HANK output and structural VARs alongside.

1.2 HANK — Heterogeneous Agent New Keynesian

The HANK revolution began with Kaplan, Moll, and Violante's "Monetary Policy According to HANK" (AER 2018). The motivation: in the standard NK model, monetary transmission runs primarily through intertemporal substitution — when the central bank cuts rates, the representative consumer brings consumption forward. But the data show that most households have very small intertemporal elasticities. Many households hold low liquid wealth and have high marginal propensities to consume (MPCs) out of transitory income — they are effectively "hand-to-mouth," even if they hold substantial illiquid wealth (a house, retirement account).

KMV combine uninsurable idiosyncratic income shocks with two assets (a liquid asset and an illiquid one) to generate empirically realistic wealth distributions and MPCs. In their model, an interest-rate cut works mostly via indirect channels: lower rates raise aggregate demand, which raises labor demand, which raises labor income of the hand-to-mouth, who then spend most of it. The direct intertemporal channel — the workhorse of RANK (Representative Agent NK) — is small (Kaplan, Moll, Violante AER 2018).

This reorders three policy intuitions:

1. Fiscal policy matters more for monetary transmission than RANK suggested. When the Fed cuts and Treasury borrowing costs fall, the fiscal authority's response (does it spend? rebate? consolidate?) becomes a first-order determinant of the macro response.
2. Distributional effects are not just normative — they are positive. Where the income from a rate cut accrues (capital owners vs wage-earners) shapes the size of the aggregate response.
3. Forward guidance is weaker than RANK predicts. McKay–Nakamura–Steinsson (AER 2016, "The Power of Forward Guidance Revisited") show that introducing borrowing constraints and uninsurable risk substantially blunts the response of current consumption to far-future rate promises — a partial resolution of the forward-guidance puzzle (McKay, Nakamura, Steinsson AER 2016).

The technical bottleneck for HANK was computational: solving a model with a full wealth distribution as a state variable was expensive. Auclert, Bardóczy, Rognlie, and Straub's "Using the Sequence-Space Jacobian" (Econometrica 2021) made HANK tractable for estimation and policy analysis (Auclert et al. Econometrica 2021). The sequence-space Jacobian computes derivatives of perfect-foresight aggregate paths around the steady state and combines them via a linear system to solve the model — vastly faster than traditional state-space perturbation. Their open-source Python toolkit is now standard infrastructure (shade-econ/sequence-jacobian GitHub).

Study this next. To work seriously with HANK you need: (a) dynamic programming for the household problem; (b) measure theory to handle the distribution of wealth as a state; (c) perturbation methods for linearization around steady state; (d) functional analysis for sequence-space methods. Standard references: Ljungqvist–Sargent Recursive Macroeconomic Theory (4th ed., 2018); Adda–Cooper Dynamic Economics (2003); Stokey–Lucas Recursive Methods in Economic Dynamics (1989) for the math; Auclert, Bardóczy, Rognlie & Straub's online appendices for sequence-space tooling.

A second major HANK result: Auclert–Rognlie–Straub's "The Intertemporal Keynesian Cross" (JPE 2024). Here the key new object is the intertemporal MPC (iMPC) — the entire impulse response of household consumption to a transitory income shock, not just the contemporaneous response. iMPC is a sufficient statistic for fiscal multipliers in their framework. Their empirical work (using Norwegian lottery data) shows that MPCs out of unexpected income are high in the first year (~0.5) and remain elevated for several years — patterns that representative-agent and even simple TANK models cannot match, but that richer HANK models can (Auclert, Rognlie, Straub JPE 2024).

1.3 TANK — the two-agent stepping stone

The Two-Agent New Keynesian (TANK) model, due primarily to Florin Bilbiie ("The New Keynesian Cross," JME 2020 and prior work), is a stripped-down stand-in for HANK. The economy has just two household types: a fraction λ of hand-to-mouth (rule-of-thumb) consumers who spend all their current income, and a fraction (1−λ) of unconstrained optimizers who behave as in standard NK.

Crucially, what generates HANK-like amplification is not just the existence of hand-to-mouth agents but the distributional rule that gives them income that rises more than one-for-one with aggregate income (cyclical income inequality). When that holds, the "Keynesian cross" — slope of planned expenditure with respect to income — exceeds one, generating large multipliers and amplification (Bilbiie JME 2020; Debortoli–Galí 2020 TANK working paper).

Why TANK matters pedagogically. TANK is the algebra you can do on paper. It captures roughly the right qualitative shape of HANK predictions for things like fiscal multipliers and the strength of indirect monetary transmission, while being amenable to closed-form solutions. The model "Macroeconomic Models for Monetary Policy" surveys treat TANK as the natural bridge from textbook NK to research-frontier HANK.

1.4 The Fiscal Theory of the Price Level (FTPL)

For most of the post-Volcker era, mainstream macro treated the price level as determined by monetary policy: a credibly inflation-targeting central bank pins down expected inflation, and the price level adjusts to clear money/credit markets. The FTPL, developed by Leeper (1991), Sims (1994), Woodford (1994, 1995), and pushed hardest by John Cochrane, inverts this. The core identity (in continuous compounding):

B_t / P_t  =  PV_t [ s_{t+1} + s_{t+2} + ... ]

where B_t is the nominal value of outstanding government debt, P_t is the price level, and s_{t+k} are real primary surpluses. If markets expect future surpluses to fall (the government has issued debt it will not — or cannot — fully back with future taxes net of spending), the equation must rebalance somewhere. In FTPL it rebalances via a rise in P_t that reduces the real value of debt to match the (lower) present value of surpluses (Cochrane, The Fiscal Theory of the Price Level, Princeton 2023; see also Cochrane's website).

For decades FTPL was contested on theoretical grounds (does the equation pin down the price level or just describe an equilibrium constraint?) and dismissed empirically because the US ran low inflation for thirty years despite mounting debt. The 2021–2022 inflation surge — coinciding with the largest peacetime fiscal expansion in US history — gave FTPL a natural experiment.

Bianchi, Faccini, and Melosi (QJE 2023, "A Fiscal Theory of Persistent Inflation") formalize a New Keynesian model with funded and unfunded fiscal shocks. A funded shock raises debt with expected future fiscal adjustment (taxes go up later); it has little inflationary effect. An unfunded shock raises debt with no expected adjustment; it generates persistent inflation that the central bank can only partly offset without destroying the recovery. Their estimated model attributes a significant share of 2021–2022 inflation to unfunded fiscal expansion (Bianchi, Faccini, Melosi NBER 30727; published QJE 2023).

The FTPL is contested on three fronts:

• Empirically. Reis and others argue inflation expectations remained well-anchored, which is inconsistent with a fiscal-dominance story where households expect debt to be inflated away. Bernanke–Blanchard's 11-country decomposition gives much more weight to commodity and sectoral supply shocks than to fiscal channels (see §2).
• Theoretically. Critics argue FTPL conflates an equilibrium identity with a determination theory; the "unfunded" / "funded" distinction depends on unobservable expectations.
• Politically. If FTPL is right, fiscal policy is the senior partner — central bank independence cannot deliver price stability when the fiscal authority is not "Ricardian." This has uncomfortable implications, especially given §6 on independence.

The Reis "How do central banks control inflation?" (LSE 2023, 2024) is the cleanest synthesis from the anti-fiscalist side (Reis "How do central banks control inflation? A guide for the perplexed"). His framework presents central bank control as three layers: control of the unit of account (a fiat money equilibrium); control of short-term real rates (interest on reserves, balance-sheet composition); and control of the fiscal-backing function via central bank solvency. He argues 2021–2022 was not a fiscal-dominance episode but a conventional inflation surge from energy, supply, and demand shocks compounded by a slow central bank reaction function.

Resolution. As of 2026 the field has not settled this. The closest to consensus: some fiscal contribution is now part of mainstream accounts (Bernanke–Blanchard 2023 acknowledge it as a tail factor); fiscal sustainability and inflation are linked in ways FTPL forced the field to take seriously; but the strong FTPL view — that the price level is fundamentally determined by fiscal expectations — remains a minority position.

1.5 RBC as the benchmark

Real Business Cycle (RBC) theory (Kydland–Prescott 1982) is rarely the working model anywhere anymore but remains the conceptual benchmark. The RBC question — "how much of the business cycle can we explain with productivity shocks in a frictionless competitive economy?" — is the null hypothesis. The historical answer (Cogley–Nason 1995; King–Rebelo 1999) was: a non-trivial share of low-frequency output variance, but very little of inflation, and poor account of labor-market and asset-price dynamics. Modern NK and HANK models embed an RBC core (the optimizing-household consumption-savings block) and add frictions (sticky prices, sticky wages, heterogeneity) to get the rest.

1.6 Where the workhorses sit in 2026 — a practical map

To orient the reader: a working monetary economist at the Fed or ECB today routinely consults all of the following frameworks in policy preparation, rather than picking one:

• The estimated medium-scale NK DSGE (Smets–Wouters for academic reference; FRB/US at the Fed; NAWM-II at the ECB). Used for the "central tendency" projections, stress tests, and counterfactual policy analysis.
• A semi-structural model (FRB/US in the US case) that is less internally consistent but more flexible for ad hoc judgment.
• A structural VAR (sVAR) identified by short-run restrictions or sign restrictions, providing a reality check on the DSGE.
• A HANK model, typically built on the sequence-space-Jacobian toolkit, used for distributional analyses and for sanity-checking the size of monetary transmission.
• Reduced-form Phillips-curve and Beveridge-curve workhorses (HHNS-style flat estimates; Benigno–Eggertsson nonlinear variants; Cerrato–Gitti regional).
• A balance-sheet flow-of-funds model (the post-2008 plumbing model, often maintained by the markets desk separately from the economics staff).
• An r-star tracker (HLW principally, with alternative estimates appended).

This pluralism is itself a finding. The post-2021 episode taught policymakers that no single framework would have given the right answer in real time. The Fed staff who built the dominant pre-2020 DSGE/forecast were the same people who, working with the same tools, badly underforecast 2021–2022 inflation. The right response was not to discard the DSGE but to lean more heavily on multi-model triangulation.

1.7 Behavioural macro and bounded rationality

A parallel but currently smaller research program imports insights from behavioural economics into macro:

• Bordalo, Gennaioli & Shleifer ("diagnostic expectations") argue that agents overreact to recent salient news, generating sentiment-driven booms and busts.
• Gabaix's "Behavioral New Keynesian Model" (AER 2020) replaces rational expectations with cognitive discounting — agents underweight distant future outcomes, dampening forward guidance and reducing the size of asset-pricing puzzles.
• Coibion–Gorodnichenko inattention. Sticky-information / rational inattention frameworks find empirical support in survey data: agents update inflation expectations sluggishly.

Behavioural macro has not yet displaced rational expectations as the modal modelling choice but is increasingly the back-pocket robustness check, especially for forward-guidance and asset-pricing applications.

2. The post-COVID inflation post-mortem (2021–2024)

The 2021–2023 inflation surge is the single most important data point in macroeconomics for a generation. US headline CPI peaked at 9.1% YoY in June 2022 (highest since 1981). Core PCE peaked at 5.6% in February 2022. Euro area HICP peaked at 10.6% YoY in October 2022. The disinflation that followed — without a US recession through 2024 and 2025 — surprised most forecasters.

2.1 The supply-vs-demand decomposition

Adam Shapiro (San Francisco Fed) developed the most widely cited PCE inflation decomposition. He classifies each disaggregated PCE category in each month as demand-driven (unexpected price and quantity move the same direction) or supply-driven (price and quantity move opposite directions); residual is "ambiguous." Implemented monthly on the SF Fed website, the decomposition shows that supply-driven inflation peaked sharply in 2021 and 2022 (commodity shocks, sectoral bottlenecks), while demand-driven inflation built more gradually and lingered into 2023 (Shapiro "Decomposing Supply and Demand Driven Inflation," FRBSF WP 2022-18; Shapiro JMCB 2025/2026; data series).

The Bernanke–Blanchard (2023, 2024) work is the most ambitious. The "What Caused the U.S. Pandemic-Era Inflation?" Hutchins Working Paper (2023) and the follow-up "Analysis of Pandemic-Era Inflation in 11 Economies" (Brookings 2024) build a simple structural model of inflation as the joint dynamics of prices, wages, short- and long-run expectations, conditional on energy, food, sectoral-shortage shocks and labor-market tightness. Estimated on US data and then extended to 10 other advanced economies, the model attributes early inflation overwhelmingly to product-market supply shocks (energy, food, sectoral) with a modest labor-market contribution, but finds that labor-market tightness became a more important factor as the surge persisted (Bernanke & Blanchard "An Analysis of Pandemic-Era Inflation in 11 Economies," Brookings WP91, May 2024).

A useful mental model. Think of inflation as a sum: π = π_supply + π_demand + π_expectations-drift. The 2021 burst was disproportionately π_supply (energy, supply chains, sectoral demand reallocation, labor reallocation costs). 2022 added a meaningful π_demand contribution (fiscal stimulus working its way through; reopening demand; pent-up savings). 2023 was the disinflation phase: π_supply collapsed as supply chains healed (the "immaculate disinflation"), labor markets cooled, while expectations never fully de-anchored.

Goolsbee at the Chicago Fed has been the most vocal Fed voice for the supply-led reading, arguing that easing supply chains and rising labor force participation explain much of the 2023 disinflation (Goolsbee "Lessons from the Supply Side," Feb 2025; Biden CEA "Chain Reaction: Immaculate Disinflation and the Role of Easing Supply Chains" 2023). The San Francisco Fed's "Global Supply Chain Pressures and U.S. Inflation" letter attributes ~60% of the 2021 inflation surge to supply-chain pressures (FRBSF Economic Letter, June 2023).

2.2 The fiscal contribution

How much of the surge was fiscal? The American Rescue Plan (March 2021) injected $1.9tn into an economy already approaching its pre-COVID output trend. Summers had warned in February 2021 that this risked overheating; Krugman, DeLong, and Treasury orthodoxy initially dismissed the concern. By mid-2021 inflation was running well above target.

Three quantitative attempts at the fiscal share:

• Bianchi–Faccini–Melosi (QJE 2023): identify a substantial unfunded-shock contribution to trend inflation in 2021–2022 in their structural model. They argue the standard NK accounting underestimates the fiscal channel because it omits unfunded shocks.
• De Soyres et al. (Fed FEDS Notes 2022): estimate the ARP added about 2–3 percentage points to US inflation in 2021–2022 vs counterfactual. A more demand-side, conventional accounting.
• Cochrane (2022, 2023): treats the entire surge as fundamentally fiscal in origin, with monetary policy responding too slowly to prevent the fiscal repricing.

The fiscal/non-fiscal share remains contested. The middle-of-the-road read circa 2026: fiscal expansion was a meaningful contributor (maybe 25–40% of the surge); supply and reopening shocks were larger; the central bank's late reaction function (the Fed held rates at zero into March 2022) compounded both.

2.3 The "greedflation" / sellers'-inflation debate

Isabella Weber (UMass) and Evan Wasner argued in "Sellers' Inflation, Profits and Conflict" (2023) that firms with market power exploited the supply-shock environment as a coordination device — the supply shock provided a focal point for simultaneous price increases that no individual firm could otherwise risk. Mike Konczal and Niko Lusiani's Roosevelt Institute analysis of 3,698 US firms showed 2021 markups at their highest level since the 1950s (Konczal & Lusiani, "Prices, Profits, and Power" 2022; Weber & Wasner "Sellers' Inflation" Project Syndicate 2023).

Mainstream economists pushed back. The complaint: markup expansion in a supply-shock environment is exactly what economic theory predicts when firms face inelastic short-term demand. Attributing the price increase to "greed" rather than "scarcity" is mostly a labeling exercise. Subsequent data showed markups normalizing as supply healed. By 2024 the consensus view: margins did expand during the shock, but the causal direction is ambiguous (did markups cause inflation, or did inflation provide cover for markups?), and the empirical contribution to aggregate inflation was likely modest, perhaps 10–20% of the surge.

Tariffs in 2025–2026 reopened the question. Brian Albrecht and others argued that the 2025 Trump tariffs would generate "tariff-driven greedflation" — passing through more than 100% of cost — if firms used the cover (Albrecht "Will tariffs cause greedflation?" Economic Forces).

2.4 What we learned about inflation expectations

The most surprising — and most encouraging — feature of the 2021–2024 episode: medium-to-long-term inflation expectations remained anchored. University of Michigan 5–10 year expected inflation peaked at about 3.1%, well below realized inflation. Market-based 5y5y forward inflation breakevens stayed close to 2.5% throughout. Survey of Professional Forecasters long-run expectations barely moved.

Coibion, Gorodnichenko, and coauthors documented heterogeneity in household and firm expectations and found that household inflation expectations are noisy, drift upward when inflation surges, but respond meaningfully to central bank communication. Firms in their JME 2024 paper showed substantial pass-through of expected inflation into wage-setting and hiring decisions, validating a key NK transmission channel (Coibion, Gorodnichenko, IZA DP 17919, 2025; Candia, Coibion, Gorodnichenko, "Inflation Expectations of U.S. Firms").

The bottom line: well-anchored expectations did most of the disinflationary work, vindicating both the rational-expectations and the credibility-of-policy frameworks of the post-Volcker era. This is one of the very few unambiguous "settled" findings.

2.5 Timeline of the surge and the disinflation

A compressed timeline is useful because dates matter for the interpretive debates:

• March 2020. Fed cuts to ZLB; QE restart at unlimited pace; CARES Act ($2.2tn) passed.
• April 2020. Headline US CPI falls to 0.3% YoY as activity collapses.
• December 2020. Consolidated Appropriations Act ($900bn).
• March 2021. American Rescue Plan ($1.9tn). Headline CPI: 2.6% YoY.
• April 2021. First "transitory" warnings; Summers' op-ed in the Washington Post predicts overheating.
• June 2021. Headline CPI: 5.4%. Fed's June SEP raises 2023 rate forecast modestly.
• November 2021. Powell renominated; "retire the word transitory" in testimony.
• March 2022. Fed begins liftoff (+25bp). Headline CPI: 8.5%. Russia invades Ukraine; oil and gas spike.
• June 2022. Headline CPI peak: 9.1%. Fed delivers first +75bp hike. QT begins.
• November 2022. Fed delivers fourth consecutive +75bp hike. Headline CPI: 7.1%, moderating.
• March 2023. SVB, Signature, First Republic failures. Fed continues hiking despite banking stress, with BTFP/discount window support for banks.
• July 2023. Fed Funds peak at 5.25–5.50%. Disinflation visibly underway.
• September 2023. Fed pauses. Long end yields spike to 5%+.
• March 2024. BoJ exits YCC; raises policy rate above zero.
• September 2024. Fed begins cutting (50bp).
• December 2024. Fed Funds at 4.25–4.50%.
• August 2025. Powell at Jackson Hole announces FIT framework, opens door to September cut.
• September 2025. Fed cuts to 4.00–4.25%.
• October–December 2025. QT ends (effective Dec 1, 2025); SRF becomes SRP with full allotment.
• Through 2026. Disinflation continues; core PCE in mid-2025 still elevated (~2.9%); Fed framework stress-tested by Trump administration's pressure on independence.

The pattern: the Fed was visibly behind the curve from mid-2021 to March 2022 (nine months of inflation above 5% with policy rate at zero); aggressive once it moved; helped substantially by supply healing in 2023; and managed the SVB crisis without abandoning the disinflation campaign. Whether this should be remembered as a triumph or a series of lucky breaks depends on whom you ask.

3. The Phillips curve: dead, flat, nonlinear, or just hibernating?

The Phillips curve — the negative relationship between unemployment and inflation, named for A.W. Phillips' 1958 paper on UK wage inflation — has had multiple obituaries written. The pre-2020 view: post-1990 the relationship was essentially flat (a 1pp fall in unemployment delivered only ~0.1pp of inflation), and the curve had stopped being a useful policy guide.

3.1 The "very flat" finding

Hazell, Herreño, Nakamura, and Steinsson (QJE 2022) used cross-state US variation to identify the slope of the Phillips curve. Their key finding: the curve has always been very flat, even in the early 1980s. The dramatic fall in inflation under Volcker was driven primarily by a downward shift in long-run inflation expectations, not by Volcker-induced unemployment moving the economy along a steep curve (Hazell, Herreño, Nakamura & Steinsson QJE 2022, 137(3): 1299–1344; NBER WP 28005). This was a major intellectual event: it implied that anchoring expectations does almost all the heavy lifting, and that the curve itself is a weak structural relation.

3.2 The nonlinearity revival

The 2021–2023 surge made the "flat curve" view harder to maintain. With unemployment falling to 3.4% and inflation spiking, the inflation/unemployment relation looked anything but flat at the right tail.

Benigno and Eggertsson's "It's Baaack: The Surge in Inflation in the 2020s and the Return of the Non-Linear Phillips Curve" (NBER WP 31197, 2023; AEA P&P 2024) re-formalize the relation as a "slanted-L" Phillips curve: roughly flat when unemployment is high and slack abundant, steeply rising when the labor market is tight (Benigno & Eggertsson revised draft, December 2024). They use the vacancy-to-unemployment (V/U) ratio as the relevant measure of tightness rather than the unemployment gap, arguing that in a search-and-matching framework V/U is the relevant measure of marginal hiring cost.

Cerrato and Gitti (2022, building on a Gitti Cleveland Fed paper) find using US metropolitan-area panel data that the post-COVID regional Phillips curve was three times steeper than pre-COVID (Cerrato & Gitti, "Nonlinearities in the Regional Phillips Curve with Labor Market Tightness," Cleveland Fed conference paper 2024). The Czech National Bank's research staff (CNB "The Phillips curve: A new era of non-linearity?") and the ECB's WP 3133 on the euro area price Phillips curve (ECB WP 3133) reach similar nonlinear conclusions.

3.3 The skeptics

Not everyone is convinced. Hazell, Patterson, and others published "On the Fragility of the Nonlinear Phillips Curve Interpretation of Recent Inflation" (NBER WP 33522, 2025) showing the nonlinear finding is sensitive to specification choices — particularly the choice of trend/cycle decomposition and the treatment of supply-shock-driven inflation (NBER WP 33522; CEPR VoxEU summary). The intuition: if you control properly for the supply shocks, the residual price-wage-unemployment relation looks much closer to the pre-2020 flat curve.

3.4 The "slope of slacks" view

A reconciling framework due to Benigno–Eggertsson and the Boston Fed labor staff: the Phillips curve is convex in the V/U ratio. When unemployment is high and V/U is low, you can absorb additional demand with little inflation. When V/U passes a threshold (perhaps around 1.0 in US data; the COVID peak was ~2.0), the curve steepens dramatically. This rationalizes both the HHNS flat-curve finding (estimated over a period dominated by slack labor markets) and the post-COVID steep-curve finding.

For policy: this means anchoring expectations is necessary but not sufficient. A central bank that allows the labor market to overshoot the "slope cliff" can generate inflation that does not require a de-anchoring shock to be persistent.

Algebra — see §6.1. The standard linearized NK Phillips curve and its nonlinear extensions appear in the Mathematical Foundations section.

3.5 What it means for forecasters

The practical takeaway from the Phillips curve literature for someone trying to forecast inflation:

• Anchored expectations are the dominant determinant of medium-term inflation. The "anchor" is a credibility variable: it can shift with policy actions or with prolonged target deviations.
• The output gap matters, but its coefficient is small in slack conditions. Don't expect a 1-point output gap to produce 1 point of inflation.
• When labor markets enter the "tight zone" (V/U above ~1; unemployment well below NAIRU), the curve becomes more pronounced and reactive — and supply shocks become more inflationary, because firms have less unused capacity to absorb shocks via inventory rather than prices.
• Sectoral shocks can produce headline inflation that core inflation does not capture. The 2021 episode was partly missed because the Fed focused on core and missed energy/auto sectoral dynamics that were spilling over.
• Survey expectations matter both as inputs into pricing and wage decisions and as a thermometer of de-anchoring risk. Watching the 5y5y forward TIPS breakeven, the SPF long-run consensus, and the Michigan survey 5-10y measure together gives a more reliable picture than any one of them.

4. Monetary policy frameworks (2020–2026)

The 2020–2022 period saw the most fundamental rethinking of monetary policy frameworks since the 1990s inflation-targeting wave. The 2024–2026 period saw a partial reversal.

4.1 The Fed: from FAIT (2020) to FIT (2025)

In August 2020 the Fed adopted Flexible Average Inflation Targeting (FAIT). The motivation was the post-GFC experience of chronically below-target inflation and the binding ZLB. FAIT was asymmetric: after periods of inflation below 2%, the Fed would aim for inflation moderately above 2% "for some time" to make up the shortfall. The employment side dropped the symmetric "deviations from maximum employment" language and replaced it with attention to "shortfalls" — implicitly accepting that very tight labor markets need not trigger preemptive tightening.

FAIT was designed for the world the Fed thought it inhabited in 2019: secular stagnation, binding ZLB, chronically undershoot inflation. The 2021–2022 inflation surge made it look ill-suited to symmetric risks. Two specific problems:

1. Make-up inflation became binding too late. FAIT was triggered by under-target periods before 2021, but the Fed couldn't credibly say in mid-2021 that running 4–5% inflation was "make-up" — it was overshooting.
2. The asymmetric employment language created a presumption against preemptive hiking. It made the Fed appear behind the curve.

In August 2025 Powell announced at Jackson Hole the new framework: Flexible Inflation Targeting (FIT), essentially a return to the 2012 framework with refinements. The key changes (Powell speech Jackson Hole, Aug 22, 2025; FOMC framework statement; Brookings analysis):

• Removed all "averages 2 percent over time" language. The target is 2% on the symmetric, contemporaneous PCE inflation rate.
• Removed "shortfalls" from the employment language; replaced with attention to "deviations" in either direction, restoring symmetry.
• De-emphasized the ZLB as a defining feature of the policy environment; the statement now references "a broad range of economic conditions."
• Preserved the dual-mandate framework with appropriate flexibility about which mandate dominates in non-complementary situations.

Powell's Jackson Hole speech also opened the door to a September 2025 rate cut, marking the start of an easing cycle.

The Fed's FEDS Note "A Roadmap for the Federal Reserve's 2025 Review" provides the most complete official rationale.

4.2 The ECB: 2021 review and 2025 reassessment

The ECB's first strategy review since 2003 concluded in July 2021. It made the inflation target symmetric at 2% (replacing "below, but close to, 2%"), explicitly endorsed quantitative tools as standard policy at the ELB, and committed to incorporating owner-occupied housing into the HICP measurement over time (ECB July 2021 press release).

The follow-up 2025 strategy assessment (released June 30, 2025) confirmed the symmetric target and added an emphasis on managing a more uncertain, more inflation-volatile world. The next assessment is scheduled for 2030 (ECB strategy assessment overview, June 2025; ECB press release June 30, 2025; Lagarde speech "Strategy assessment: lessons learned").

4.3 The Bank of England

The Bank of England maintained its 2% CPI target throughout, but the 2022–2024 period saw real-time controversy over the speed of tightening. The BoE under Bailey was slower to begin hiking than the Fed; UK inflation peaked higher (11.1% YoY in October 2022) and disinflation has been slower. The BoE has not undertaken a formal framework review on the scale of the Fed or ECB.

The BoE's most significant policy issue post-2024 is the Asset Purchase Facility (APF) losses indemnified by HM Treasury. The APF was created in 2009 with a Treasury indemnity covering any APF losses. As Bank Rate rose well above the average yield on APF holdings, the APF began running large losses. Between March 2023 and February 2024, the Treasury transferred £44.5 billion to the BoE to cover APF losses (Investment Week, "Government funnels almost £45bn to Bank of England"; CETEx "The Bank of England's Asset Purchase Facility"). Lifetime APF losses are now projected at £150bn+ by the early 2030s. This is a quasi-fiscal cost of QE that was not seriously priced into the original asset-purchase decisions — a real lesson for future balance-sheet policy.

4.4 The Bank of Japan: the end of YCC

The BoJ's exit from its decade-long ultra-accommodative regime is the most dramatic regime change of the period. The BoJ had operated:

• A policy rate of −0.1% since 2016 (the world's only sustained negative policy rate at the time).
• Yield Curve Control (YCC) since 2016: targeting 10-year JGB yields at 0%, with a tolerance band that widened from ±10bp through ±0.5% to ±1.0% by late 2023.
• A massive ETF and JGB purchase program.

In March 2024, Governor Ueda's BoJ:

• Raised the policy rate range to 0% to 0.1% (first rate increase in 17 years).
• Ended Yield Curve Control.
• Ceased ETF and J-REIT purchases.
• Indicated continued JGB purchases at roughly the prevailing pace, but with the YCC framework removed.

(Bank of Japan, "Changes in the Monetary Policy Framework," March 19, 2024; CNBC coverage)

The rationale Ueda gave: the likelihood of stable 2% inflation had crossed a threshold, after decades of failing to exit the deflation/lowflation trap. The exit has been gradual; further rate hikes through 2025 brought the policy rate to ~0.5%. The JGB yield curve has steepened materially, and yen volatility has risen.

The lesson from the YCC exit is being absorbed in real time: a credible YCC commitment requires the central bank to absorb essentially unlimited JGB supply at the target yield, which becomes increasingly costly when inflation expectations rise. Exit must be both rapid (to avoid being overwhelmed by speculative attack) and well-signaled (to avoid market dysfunction). The BoJ managed both reasonably well.

4.5 Make-up strategies vs forward guidance

Reifschneider–Williams (2000) is the foundational analysis of the "lower for longer" strategy at the ELB. They argued that during a ZLB episode the central bank should track cumulative deviations of the policy rate from a Taylor-rule prescription, then keep rates lower for longer than the rule would dictate once the ZLB no longer binds, to make up the lost monetary stimulus (Reifschneider & Williams, Boston Fed conference paper).

This is the intellectual origin of FAIT's "make-up" idea. The post-FAIT consensus: make-up strategies are theoretically sound at the ELB but politically/operationally hard to commit to credibly once exiting. The Fed's 2025 revision essentially conceded that FAIT did not survive contact with the symmetric inflation environment.

4.6 How the major central banks differ — at a glance

4.7 The "make-up vs preemption" debate

A key intellectual divide in monetary policy design:

• Make-up strategies (FAIT in original form; price-level targeting in some proposals): when inflation undershoots, you promise to overshoot later. Theoretically powerful at the ELB because the credible promise of future overshoot raises current inflation expectations and lowers real rates. Practically hard to sustain — central banks face strong asymmetric pressure to tighten on overshoot but loose commitment to "overshoot back" after an undershoot.
• Preemptive tightening (the pre-2020 norm): when unemployment falls below NAIRU, hike before inflation appears, on the logic that the Phillips curve will eventually deliver inflation if you wait too long.
• Outcome-dependent (the Fed's current FIT): more or less, "we'll move when we see actual deviations" — which sacrifices preemption but reduces the risk of acting on noisy potential-output estimates.

The 2021–2022 episode was a setback for the original FAIT but not necessarily for preemption: a strict Taylor rule responding aggressively to early CPI readings would have started tightening in mid-2021. Whether the Fed's 2025 FIT framework will be sufficiently preemptive in the next cycle is an open question.

5. Balance-sheet policy: the new permanent regime

QE began as crisis-mode unconventional policy in 2008. By 2024–2026 it had become a permanent operating mode of monetary policy in advanced economies. Central banks now operate in "ample reserves" regimes where the policy rate is steered through administered rates rather than scarcity of reserves.

5.1 The Fed's framework and recent changes

The Fed's modern operating framework uses three main administered rates:

• IORB (Interest on Reserve Balances). The rate the Fed pays banks on reserves held at the Fed. Effective floor for the Fed Funds rate.
• ON RRP (Overnight Reverse Repo). A facility allowing money market funds and other non-bank counterparties to lend to the Fed overnight at a pre-set rate, providing a floor for non-bank money market rates.
• SRF/SRP (Standing Repo Facility, recently renamed Standing Repo Program). A backstop where primary dealers and certain banks can borrow from the Fed against Treasuries at a pre-set rate, capping spikes.

The asset side of the balance sheet expanded from ~$900bn pre-2008 to a peak of ~$8.9tn in 2022. Quantitative Tightening (QT) began in June 2022 — letting maturing securities roll off rather than reinvesting. The Fed slowed QT in June 2024 (from $60bn/month to $25bn/month in Treasuries; MBS pace unchanged) and ended QT effective December 1, 2025 (FOMC announcement October 2025; Cleveland Fed "QT, Ample Reserves, and the Changing Fed Balance Sheet"). The peak-to-trough decline was about $2.2 trillion.

Post-QT, the Fed will fully reinvest maturing Treasuries (with a $5bn/month cap currently) and continue MBS runoff at $15–20bn/month. Reserves had fallen to about $3 trillion and are expected to settle near $2.7 trillion.

The December 2025 FOMC meeting also restructured the SRF/SRP:

• Renamed SRF → SRP (Standing Repo Program), partly to combat stigma.
• Eliminated the $500bn aggregate daily cap on SRP operations.
• Moved to "full allotment": all bids will be awarded in full.

(New York Fed Teller Window article on SRP)

Why this matters: the Fed is institutionalizing the SRP as a routine plumbing tool, similar to the ECB's regular repo operations. This represents a definitive break from the pre-2008 corridor system and a doubling-down on the floor system.

5.2 The reverse repo experiment

The ON RRP was relatively small for most of its 2013–2020 existence. Post-COVID it ballooned: as fiscal transfers swelled bank deposits and money market funds were flooded with cash, the ON RRP balance peaked above $2.5 trillion in late 2022 — accounting for roughly half of money market fund assets (Federal Reserve "Bank Deposit Flows to Money Market Funds and ON RRP Usage").

The ON RRP balance has fallen sharply since 2024 as money market funds extended maturities to lock in pre-cut yields and as the Fed's QT drained reserves. By late 2025 ON RRP usage was a fraction of its peak. This implicit substitution between ON RRP and reserves is now a central feature of the Fed's plumbing — and a key uncertainty in calibrating what level of reserves counts as "ample."

5.3 Central bank capital and remittances

QE generated large central bank profits when long-dated assets paid yields above the Fed's near-zero interest expense on reserves. Those profits flowed to Treasury (the Fed remitted $80–100bn/year through 2021). The reversal — as short rates rose well above the average asset yield — produced large operating losses.

The Fed handles operating losses via a "deferred asset" on its balance sheet (an accumulating IOU against future remittances). The accumulated Fed deferred asset reached ~$220bn by 2026 (PIIE "How much money have central banks really lost?"). The Fed will not remit again to Treasury until that asset is fully worked off via future profits.

The Bank of England — operating under explicit Treasury indemnity — saw direct fiscal transfers of £44.5bn (March 2023 to February 2024). The ECB, with no comparable indemnity but with national central banks bearing losses, saw the Bundesbank and others post historically large losses absorbed against capital and provisions.

Why this matters: central bank losses are not just accounting — they shape the politics of independence. A central bank that prints to pay interest on reserves while requesting fiscal transfers is in a more precarious position vis-à-vis the legislature than one steadily remitting profits.

5.4 Does QE work? The empirical state

The Krishnamurthy–Vissing-Jorgensen and Gagnon–Raskin–Remache event studies established that QE announcements moved long-term yields by tens of basis points. The signaling channel (commitment to keep rates low) and the portfolio-balance / preferred-habitat channel (reducing duration supply forces investors into riskier assets) are both empirically supported.

The newer debate: does QE work in non-crisis conditions? The marginal asset purchases of 2020–2021 were unlikely to be lubricating dysfunctional markets the way the 2008–2009 purchases were. The 2022 reversal of QE produced much smaller adverse effects than QE had positive effects — a sign of state-dependence: QE is most powerful when markets are dysfunctional, much less powerful when they are not.

This implies an asymmetry that should make central banks cautious about routine QE. QT, by contrast, looks relatively benign in calm markets and dangerous in stressed ones (cf the September 2019 repo spike and the March 2020 Treasury market dysfunction).

5.5 The September 2019 repo episode — a cautionary tale

In September 2019, the US repo market suddenly seized up. Overnight repo rates spiked from ~2% to as high as 10% over two days. The Fed had to inject hundreds of billions in temporary repo operations to stabilize rates and ultimately to begin "non-QE" Treasury bill purchases.

What went wrong: the Fed had been running QT for two years, draining reserves from a 2017 peak of ~$2.8 trillion to ~$1.4 trillion by September 2019. Reserves had fallen below the level banks were comfortable with, leaving the system short of "scarce reserves" precisely when corporate tax payments and Treasury settlements drained liquidity simultaneously.

This episode taught the Fed:

• "Ample" reserves are higher than they look. The Fed thought it could run reserves down further; it was wrong.
• The Standing Repo Facility (created in 2021) was a direct response — a permanent backstop so that a recurrence would be impossible.
• The 2022–2025 QT run was much more carefully calibrated, slowing in 2024 and ending in December 2025 with reserves well above the September 2019 trough.

The episode is a recurring reference in the current "what is ample?" debate. The Fed's caution about ending QT too aggressively can be traced directly to September 2019.

5.6 The March 2020 Treasury market dysfunction

In March 2020, as COVID lockdowns hit and global investors raced for cash, the US Treasury market — supposedly the world's deepest and most liquid — broke down. Bid-ask spreads on 10-year notes widened five-fold; high-grade corporate bond markets seized; foreign central banks were forced sellers of Treasuries to raise dollar liquidity.

The Fed responded with:

• Unlimited Treasury purchases ($75bn/day in the first weeks; eventually ~$1.5tn of net Treasury purchases over weeks).
• Standing dollar swap lines expanded to additional central banks; FIMA (Foreign and International Monetary Authorities) repo facility created.
• Money market mutual fund and corporate bond purchase facilities (under Treasury credit support).
• A repo program of unprecedented size.

The episode revealed that the dealer-intermediated Treasury market has a much lower capacity to absorb large-scale forced selling than pre-2020 thinking assumed. Subsequent reforms (the all-to-all trading expansion; central clearing pilot programs; the Treasury Market Practices Group recommendations) have aimed to add resilience. The standing dollar swap lines and FIMA repo are now permanent fixtures of the global financial architecture.

6. The Effective Lower Bound and r-star debate

6.1 The ELB / ZLB framework

The Effective Lower Bound (ELB) — slightly above the strictly Zero Lower Bound (ZLB) because of physical cash hoarding costs and reserve-system frictions — was the dominant constraint on monetary policy through 2009–2021. It is the primary motivation for QE, forward guidance, FAIT-style make-up strategies, and (in the European case) negative interest rate policies.

Reifschneider–Williams (2000) is the foundational analysis (see §4.5). The post-GFC ELB literature is dominated by Eggertsson, Krugman, Woodford, Williams, and the Fed staff working in Reifschneider's lineage.

6.2 Secular stagnation

Lawrence Summers' 2013 "secular stagnation" hypothesis revived Alvin Hansen's 1938 term: the equilibrium real rate that balances saving and investment at full employment may be persistently negative, making the ELB the modal policy environment.

Eggertsson, Mehrotra, and Robbins formalized secular stagnation in an overlapping-generations (OLG) NK model ("A Model of Secular Stagnation," AEJ:Macro 2019). In a 56-period quantitative life-cycle model, demographic decline and rising income inequality produce a natural rate of −1.5% to −2% — implying frequent ZLB episodes for the foreseeable future (Eggertsson, Mehrotra, Robbins AEJ:Macro 2019).

Summers and Blanchard publicly debated the future of secular stagnation in March 2023 at PIIE (transcript). Summers argued that secular stagnation was over (he saw a higher-rate, higher-inflation regime ahead); Blanchard argued the underlying forces (aging, savings glut, scarcity of safe assets) remained intact.

Subsequent developments split the difference. Summers updated to a "stagflation-flavored" view in 2024, expecting low growth and elevated inflation. Blanchard's Fiscal Policy under Low Interest Rates (MIT Press, January 2023; revised edition) maintained that fiscal policy room has expanded with persistently low r-star (Blanchard, Fiscal Policy under Low Interest Rates, MIT Press 2023; PIIE summary).

6.3 r-star estimates

The Holston–Laubach–Williams (HLW) state-space model is the most-cited r-star estimate. It uses a Kalman filter to extract the unobserved natural rate from observed real rates, output, and inflation, jointly with a measure of the trend growth rate.

Study this next. The HLW model is built on state-space modeling with Kalman filtering. To work with it you need: linear time-series analysis, the basics of Kalman filtering (Harvey 1989 Forecasting, Structural Time Series Models and the Kalman Filter), and an appreciation of why r-star is fundamentally unobservable (so the filter is doing real inference). The intuition: HLW assumes r-star follows a random walk, identifies it through its co-movement with output and inflation, and uses the joint structure to filter signal from noise.

The latest published HLW estimates (mid-2025 vintage) put US real r-star at about 0.84%, implying nominal neutral around 2.84% (New York Fed "Measuring the Natural Rate of Interest"). NY Fed President Williams in July 2024 publicly defended a "low r-star regime endures" view (Williams speech, July 2024; Bloomberg coverage).

But there is real disagreement:

• Richmond Fed staff find r-star above 2% post-pandemic (Richmond Fed Brief 24-36).
• Survey-based measures (Mercatus survey, Blue Chip) put nominal neutral closer to 3.5–4%.
• Market-implied measures (5y5y real forwards) rose substantially post-2022 and have not retraced fully.
• Holston–Laubach–Williams for the euro area suggests modest rise from pre-COVID lows (San Francisco Fed letter, May 2025).

The Bank of Finland Bulletin's modified HLW analysis (Bofbulletin 2024) and the Liberty Street post-pandemic global r* analysis (Liberty Street, February 2026) form the best ongoing real-time discussion.

Why r-star matters for policy. If r-star is 0.5%, current US real rates around 2% are highly restrictive. If r-star is 2%, current rates are barely restrictive. The Fed's "restrictive enough?" deliberations through 2024–2026 turn substantially on this estimate, and the persistent disagreement is a major source of policy uncertainty.

6.4 Why r-star estimates disagree

The structural reason for the persistent disagreement on r-star is worth understanding:

• HLW is a state-space model that treats r-star as a random walk and identifies it through the joint behavior of GDP, inflation, and real rates. It downweights short-term moves and gives a smoothed, slow-changing estimate. It is conservative on revisions, which is why post-2022 estimates remain low.
• Richmond Fed's modified HLW uses different specifications of the trend-growth process and the COVID-era treatment of distortions. The differences in modeling choices generate point estimates 100+ basis points apart.
• Survey-based measures ask economists or market participants directly for their estimate. These respond fast to recent data but encode whatever priors the respondents hold.
• Market-implied measures extract r-star from the long end of the real yield curve (e.g., 5y5y forward TIPS yields minus a term premium estimate). These move with bond market sentiment and term-premium estimates that are themselves contested.
• Macro-DSGE-implied r-star (the Laubach 2023 update; HLW vs Lubik–Matthes variants) generates yet another distribution of estimates.

A reasonable consumer of these estimates should:

1. Watch the trend not the level: all measures have moved up modestly since 2022, suggesting at least some structural change.
2. Treat the survey/market measures as expressing optimism about the future neutral rate; treat the structural-model measures as reflecting backward-looking realized fundamentals.
3. Recognize that for short-horizon policy questions, the near-term equilibrium rate (sometimes called r-star-short or r**) matters more than the structural long-run rate, and these can diverge meaningfully.

The Brookings explainer (Hutchins Center "What is the neutral rate of interest?") is the cleanest pedagogical introduction.

7. Fiscal policy: multipliers, sustainability, MMT

7.1 Fiscal multipliers

Ramey's JEP 2019 survey ("Ten Years After the Financial Crisis: What Have We Learned from the Renaissance in Fiscal Research?") is the standard reference (Ramey JEP 2019). Her synthesis:

• Government spending multipliers are usually in the 0.6 to 1.0 range during normal times.
• They may be larger in deep recessions (some estimates above 1.5), especially when monetary policy is constrained.
• Tax cut multipliers can be 2 to 3 in some specifications, though the upper end is sensitive to identification.
• Cross-state / cross-region (geographic) multipliers are typically estimated higher, with Chodorow-Reich's (AEJ:Policy 2019) careful synthesis pointing to a preferred geographic multiplier of about 1.8 (Chodorow-Reich AEJ:Policy 2019).

The geographic-aggregate mapping is subtle. A geographic multiplier of 1.8 implies — under standard assumptions about monetary offsets and cross-region spillovers — a national no-monetary-response multiplier of 1.7 or above. With an active Taylor-rule-following central bank, the national multiplier is mechanically smaller, often 0.5 to 1.

The 2020–2021 fiscal experiments provide a unique natural experiment. The CARES Act (March 2020, $2.2tn) was largely automatic stabilizer-like and prevented a much deeper recession. The American Rescue Plan (March 2021, $1.9tn) was applied to an economy approaching trend output; multiplier estimates for ARP-induced output expansion vs ARP-induced inflation are still being debated (see §2.2).

7.2 Debt sustainability and R minus G

Blanchard's 2019 AEA Presidential Address ("Public Debt and Low Interest Rates," AER 2019) reframed sovereign-debt sustainability (Blanchard AER 2019; PIIE WP 19-4). His core observation: when the nominal interest rate on government debt (r) is below the nominal growth rate of the economy (g), running a balanced primary budget produces a declining debt-to-GDP ratio. The "snowball" works in your favor.

For most of US history, g has been higher than r — the snowball is benign. Post-GFC, the (r − g) gap was strongly negative. This implies:

• Fiscal costs of debt are low when r < g, because the government rolls over debt at rates below GDP growth.
• Welfare costs of debt are also lower than typically assumed, since the implicit interest service is small.
• This is not a "free lunch" — Blanchard explicitly distinguishes from MMT — but it expands the responsible fiscal space.

The 2022–2024 rate normalization narrowed and in some cases reversed the (r − g) gap. Blanchard's 2023 book updates: even with positive (r − g), the welfare costs of debt are smaller than the textbook treatment suggests, and the room for fiscal stabilization (which requires running deficits in downturns) remains larger than orthodox finance ministry projections imply.

Critics (Reis, Mauro, IMF DSA staff) note that the negative-(r−g) world is itself partly endogenous to fiscal restraint and high investor confidence; complacency about debt sustainability can be self-fulfilling in reverse. The IMF's Debt Sustainability Framework continues to treat (r − g) as a key parameter but with conservatism on its persistence.

7.3 MMT

Modern Monetary Theory (MMT), advanced by Stephanie Kelton, Randall Wray, and others, makes two key claims often conflated:

1. Operational claim (largely accepted): a sovereign-currency-issuing government cannot be forced into "default" in the technical sense — it can always create the currency to pay obligations denominated in that currency.
2. Normative claim (highly contested): because of (1), governments should target full employment and price stability through fiscal policy (with a "job guarantee"), using taxes to manage aggregate demand rather than to "finance" spending.

Greg Mankiw's "A Skeptic's Guide to Modern Monetary Theory" (AEA P&P 2020) gives the most balanced mainstream critique: MMT contains some operational truths (the central bank can always accommodate Treasury issuance in the short run), but its core policy prescriptions do not follow rigorously from its premises, particularly around the inflation constraint (Mankiw NBER WP 26650; AEA P&P 2020).

Krugman has been characteristically blunt, accusing MMT proponents of "calvinball" — moving the goalposts when challenged. Sebastian Edwards' critique focuses on MMT's underdeveloped account of inflation dynamics and its weak engagement with the IMF/balance-of-payments constraints on emerging markets.

By 2026 the MMT debate has largely died down in the academic literature. Mainstream macro has selectively absorbed some MMT observations (the operational point about sovereign currencies; the case for fiscal stabilization in deep recessions) while rejecting the broader framework.

7.4 The state-contingent multiplier

A nuance often lost in the multiplier literature: the multiplier is state-contingent. Specifically:

• In deep recessions with constrained monetary policy (ZLB or near it), the multiplier is large — possibly 1.5–2.5. Christiano, Eichenbaum, and Rebelo (2011) showed in a NK model that the multiplier on government spending at the ZLB can be substantially above unity because the central bank does not offset the fiscal expansion.
• In normal times with an active central bank, the multiplier is smaller (perhaps 0.6–1.0) because the central bank partially offsets.
• In overheating economies with constrained monetary policy (a contested category, but applicable to 2021), the multiplier on additional spending can be negative for output (substantially adding to inflation rather than real output).
• Composition matters. Transfers to low-income households (high MPC) have larger multipliers than tax cuts for high earners (low MPC). Public investment with substantial complementarity (research, infrastructure) can have especially high multipliers because it raises private investment as well.
• Duration matters. Temporary spending generates a smaller multiplier than persistent spending in a standard NK model, but TANK/HANK models show that persistent transfers can be amplified by iMPC dynamics.

7.5 Debt dynamics in the 2020s

The US federal debt held by the public rose from 79% of GDP in 2019 to 97% in 2023 to a projected 116% by 2026 (CBO baseline). The (r−g) calculus is now considerably less favorable than during Blanchard's 2019 address — nominal Treasury yields around 4–5% versus nominal GDP growth around 4%.

Three scenarios are worth tracking:

• Benign: g recovers toward 3% real (driven by AI and immigration); r-star stays moderate (Williams' "low r-star endures" view); debt stabilizes at 110–120% of GDP through gradual primary-balance improvement.
• Stressed: g stays around 2% real; r-star rises further to 3% real (Richmond Fed view); debt grows continuously even at primary-balance equilibrium, requiring substantial primary surpluses just to stabilize.
• Crisis: external or political event triggers a sudden reassessment of US fiscal credibility; term premium spikes; the Fed faces a "fiscal dominance" trap (forced to monetize or accept severe recession).

The IMF, OECD, and CBO debt-sustainability frameworks all show the US trajectory as concerning but manageable on baseline assumptions; the risk lies in the tails.

8. The heterogeneity revolution: MPCs, distribution, transmission

The most consequential intellectual development of the 2010s–2020s in macroeconomics is the explicit incorporation of household heterogeneity. The key building blocks:

8.1 Marginal Propensities to Consume (MPCs)

The standard RANK assumption of an "Euler equation" representative consumer implies a low MPC out of transitory income (something like 0.05 per quarter for a typical interest-rate / discount-rate calibration). The data show much higher MPCs — averages around 0.2–0.4 per quarter, with the bottom of the wealth distribution showing MPCs of 0.5–0.8.

Why the disconnect? A large fraction of households are "hand-to-mouth," holding very little liquid wealth even if they have substantial illiquid assets. Kaplan, Violante, and Weidner's "The Wealthy Hand-to-Mouth" (Brookings 2014) showed about 1/3 of US households fit this pattern, with about 2/3 of these being "wealthy hand-to-mouth" (high illiquid wealth, low liquid wealth — they cannot smooth consumption because their wealth is locked in a house or retirement account).

This single observation reshapes:

• Fiscal multipliers (transfers to hand-to-mouth get spent quickly).
• Monetary policy transmission (the indirect channel via labor income is more important than the direct intertemporal-substitution channel).
• Forward guidance effectiveness (less powerful than RANK implies).
• The size of the savings-glut effect (the very rich save much more than the rest; see Mian–Straub–Sufi below).

8.2 Intertemporal MPCs (iMPCs)

iMPCs extend the MPC concept to the full intertemporal response — not just the year-1 spending out of a transitory shock, but year 2, 3, 4 and beyond. Auclert, Rognlie, and Straub (JPE 2024) show iMPCs are a sufficient statistic for fiscal multipliers in their HANK framework: knowing the full path of iMPCs lets you compute the dynamic response to any government spending path, given the iMPC × deficit interaction.

The Norwegian lottery evidence (Fagereng et al. and others) shows iMPCs remain elevated for 3–4 years after a transitory income shock — substantially more persistent than RANK predicts, but less persistent than a pure liquidity-constraint model would imply. The iMPC profile is the right test for heterogeneity-richness of a macro model.

8.3 Indirect vs direct transmission of monetary policy

In RANK, an interest rate cut moves consumption almost entirely through intertemporal substitution: lower future returns make today's consumption cheaper relative to tomorrow's. In KMV's HANK, this "direct" channel is small (because the Euler equation barely binds for the hand-to-mouth, and the elasticity of intertemporal substitution is low). The "indirect" channel — rate cut → investment up → labor demand up → wages/employment up → hand-to-mouth households spend the extra income — dominates.

This shifts the locus of monetary policy from financial markets to labor markets. It also makes the cyclical income distribution (whether wages rise with output) a key determinant of monetary transmission strength.

8.4 The savings glut of the rich

Mian, Straub, and Sufi's "The Saving Glut of the Rich" (2020, 2021) document that the top 10% of the US income distribution account for 30–40% of total private saving, with saving rates that have grown faster than for the bottom 90% (Mian, Straub, Sufi NBER WP 26941; Harvard scholar). Their argument: rising top-end income inequality has been a structural force pushing r-star down (more savings chasing the same investment) and pushing household debt up (as the bottom 90% borrow from the rich to finance consumption).

This dovetails with the Sufi Jackson Hole paper on the determinants of declining r-star, and with the broader Auclert–Malmberg–Martenet–Rognlie (NBER WP 29161) demographic story (see §11).

8.5 The 2020–2021 pandemic transfer experiment

The CARES Act and ARP delivered direct transfers to American households of a magnitude unprecedented in peacetime. This generated an unintentional natural experiment for the HANK literature.

• Bottom-quartile households received transfers worth multiples of their normal annual income.
• Their consumption rose substantially within the first few quarters (consistent with high MPC).
• Their savings rate also rose (the "excess savings" of 2020–2021 totalled ~$2.1tn for US households at peak).
• The buildup was concentrated among lower-income households, with depletion timing varying by income: high-income households' excess savings persisted longest; low-income households' depleted by mid-2022 to 2023.

Bardóczy and coauthors (2024) estimate in a HANK framework that excess savings depleted within roughly three years. The pattern is broadly consistent with the iMPC profile: high initial spending, sustained elevated spending for several years, gradual return to baseline.

This episode also illuminated a subtler heterogeneity: transfers raised the flow of real spending for the bottom quartile but raised asset prices (housing, equities) for the top quartile. The distributional consequences of pandemic stimulus were thus more complex than a simple Keynesian-cross account would suggest.

9. Labor markets: Beveridge curve, vacancies, NAIRU

9.1 The Beveridge curve shift

The Beveridge curve plots the unemployment rate (U) against the vacancy rate (V). Historically a stable, downward-sloping locus, the COVID shock pushed the US dramatically outward — vacancies surged while unemployment fell more slowly than vacancies rose.

The 2022 policy debate was whether bringing vacancies down would require a large rise in unemployment (hard landing) or could be achieved with a small rise (soft landing). This was operationalized as a debate about the slope of the Beveridge curve at the relevant point.

The pessimists. Blanchard, Domash, and Summers ("Bad News for the Fed from the Beveridge Space," PIIE/NBER 2022) used the historical post-1948 US Beveridge curve and argued that the level of vacancies in early 2022 corresponded historically to sub-2% unemployment — implying the labor market was much tighter than the unemployment rate showed, and that cooling it would require a substantial rise in unemployment (Domash & Summers NBER WP 29739, "How Tight are U.S. Labor Markets?"). They concluded a soft landing was unlikely.

The optimists. Andrew Figura and Christopher Waller (FRB 2022) argued the Beveridge curve had shifted out due to pandemic-elevated job separations, and that vacancies could fall along a steep (nearly vertical) inward shift back to the curve without much unemployment increase (FEDS Notes July 2022). Benigno and Eggertsson (NBER WP 33095, 2024 "Revisiting the Phillips and Beveridge Curves: Insights from the 2020s") formalized this prediction.

The outcome. By late 2024, vacancies had fallen substantially (from the COVID peak of ~12.2 million in March 2022 to about 7 million by end-2024) while unemployment had risen modestly (from 3.4% to about 4.2%). The soft-landing camp won the empirical fight (CEPR VoxEU "Sliding safely down the Beveridge curve").

9.2 NAIRU updates

The NAIRU (Non-Accelerating Inflation Rate of Unemployment) — the unemployment rate consistent with stable inflation — has been a moving target. Pre-COVID estimates clustered around 4.0–4.5%. The post-COVID experience of 3.4–3.7% unemployment with disinflating (rather than accelerating) inflation forced revisions downward.

CBO's NAIRU estimate has moved from ~5% (2008–2012) to ~4.4% (2019) to ~4.5% (2025), with substantial uncertainty bands. The Hazell–HHNS flat-curve finding implies NAIRU is even less identified than usual: if the Phillips curve is flat, NAIRU is hard to pin down from inflation data.

The Benigno–Eggertsson nonlinear view suggests NAIRU should be thought of less as a single number than as a region where the Phillips curve transitions from flat to steep — perhaps something like 3.8–4.5% in the US.

9.3 Wage-price spirals

A major fear of 2022–2023 was a 1970s-style wage-price spiral. Alvarez, Bluedorn, Hansen, Huang, Pugacheva, and Sollaci (IMF WP 2022) built a historical database of advanced-economy episodes meeting the pattern of accelerating prices and wages over three consecutive quarters. Their startling finding: most such episodes are not followed by sustained acceleration. The 1970s spiral was the exception, not the norm (IMF WP 2022/221; IMF blog "Wage-Price Spiral Risks Still Contained").

The mechanism: a tight labor market and a supply shock can produce a one-time burst of wage and price increases, but absent a sustained de-anchoring of expectations and accommodating monetary policy, the spiral does not persist. The 2022–2024 US experience fits the modal historical episode: wages caught up to prices, then both decelerated.

Lorenzoni–Werning's "Wage Price Spirals" (2023) provided a clean theoretical model of how supply shocks can generate sustained wage-price dynamics through distributional conflict between workers and firms — but emphasized that monetary policy can break the conflict by anchoring expectations (Lorenzoni & Werning).

9.4 Quits, hires, and the "Great Resignation"

The 2021–2022 surge in quits (the "Great Resignation") and the elevated quits/hires ratio are now well-documented. JOLTS data show quits peaking at 4.5 million per month in early 2022 (~2.9% of employment), then normalizing to ~3.2 million by 2024.

The quits surge is consistent with a "high-V/U" interpretation: workers had ample outside options and used them. The subsequent decline in quits is consistent with the labor market cooling — workers no longer face as many outside options. This co-movement between vacancies and quits is one reason the Beveridge-curve "soft landing" came: it was workers reducing job-switching, not firms doing mass layoffs, that primarily cooled the labor market.

9.5 Immigration and labor supply

A surprisingly important — and politically charged — factor in the 2023–2024 disinflation was a large surge in US net immigration. CBO estimates net international migration of ~3.3 million in 2023 and similar in 2024, well above the prior 1.0 million/year trend. This expanded the labor force and helped relieve wage pressure, particularly in lower-paid services occupations.

Goolsbee and Powell both cited the immigration-driven labor supply expansion as a contributor to the soft landing. The 2025 policy reversal (tighter immigration enforcement, deportations) is a contractionary labor-supply shock; its inflation effects will play out through 2026–2027. This is a clear case where political economy and macro are interleaved.

10. Growth, productivity, and the AI question

10.1 The productivity slowdown

The "productivity puzzle" of the 2010s is that measured Total Factor Productivity (TFP) growth has been about 0.6–1.0% per year in advanced economies, well below the 1.5–2.5% norm of earlier decades. Robert Gordon's "Rise and Fall of American Growth" (2016) argued this reflects fundamental exhaustion of the great innovations of 1870–1970.

Acemoglu, Autor, and Patterson's "Bottlenecks: Sectoral Imbalances and the US Productivity Slowdown" (2023, JPE 2024) identifies a specific mechanism: productivity growth concentrated in a few sectors creates bottlenecks in others, because innovation in any sector depends on complementary innovations from input suppliers. Sector-level variance in TFP growth has risen, and this dispersion explains a substantial share of the aggregate slowdown (Acemoglu, Autor, Patterson NBER WP 31427).

10.2 The Brynjolfsson–Rock–Syverson J-curve

Brynjolfsson, Rock, and Syverson ("The Productivity J-Curve," AEJ:Macro 2021) argue that general-purpose technologies (GPTs) like AI require massive complementary investments — reorganization, training, business-process redesign — that are themselves intangible and poorly measured in national accounts (AEJ:Macro 2021). During the build-out phase, GDP and productivity growth are underestimated because the intangible investments are expensed rather than capitalized. During the harvest phase, productivity growth is overestimated as the unmeasured intangible capital pays off in measured output.

The implication: an early-stage GPT can show flat or declining measured productivity for a decade before producing a measurable burst. The 2024–2026 generative AI boom is the test case.

10.3 AI and productivity — the open question

Acemoglu's 2024 work is the most cautious mainstream forecast: he estimates generative AI / LLMs will produce about a 0.7% level increase in TFP over the next decade — a meaningful but not transformative effect. The Goldman Sachs, McKinsey, and OpenAI-affiliated researchers tend to forecast larger effects (often 1.5–7% level effects). The honest empirical answer as of 2026: we don't yet know. Several years of data will be needed to discriminate among the forecasts.

10.4 Hysteresis

Hysteresis is the idea that recessions have permanent supply-side effects: prolonged unemployment erodes human capital, reduces labor force participation, slows innovation, and lowers the trend level of output even after demand recovers.

Yagan (2019) found that workers in US states with deeper Great Recession shocks had measurably worse employment outcomes 6–8 years later, even controlling for population mobility (Yagan NBER WP 23844). Blanchard, Cerutti, and Summers (2015) found similar patterns across advanced economies.

Demand recessions scar; supply recessions don't. This is one reason why aggressive macro stabilization in deep recessions has higher expected returns than the simple short-run cost-benefit analysis suggests.

11. Demographics and r-star

11.1 The demographic-r-star link

Aging populations push r-star down through two channels:

1. Life-cycle saving. Households save more for retirement in their 40s–50s, then dis-save in retirement. A bulge of pre-retirement workers raises aggregate savings.
2. Slower labor force growth. A growing labor force raises the marginal product of capital; a shrinking one lowers it.

Eggertsson, Mehrotra, and Robbins (AEJ:Macro 2019) embed this in an OLG NK framework. Their calibration produces equilibrium real rates of −1.5% to −2% from demographic and inequality forces alone.

Auclert, Malmberg, Martenet, and Rognlie's "Demographics, Wealth, and Global Imbalances in the Twenty-First Century" (NBER WP 29161, revised through 2024) uses a sufficient-statistic approach. They predict that compositional aging effects will raise global wealth-to-GDP ratios, lower asset returns, and widen global imbalances throughout the century (Auclert et al. NBER 29161; Jackson Hole 2024 paper "Race Between Asset Supply and Asset Demand").

11.2 The Goodhart–Pradhan inversion

Charles Goodhart and Manoj Pradhan in The Great Demographic Reversal (Palgrave 2020) take the opposite long-run view. They argue:

• The deflationary forces of the past 30 years owe much to the integration of China and Eastern Europe into the global labor market (massive labor-supply expansion → wage suppression in advanced economies).
• The reverse process is now beginning: China's labor force is shrinking; aging in advanced economies reduces labor supply; deglobalization compounds the effect.
• Net effect: structurally higher inflation pressure, higher real interest rates, falling labor share inequality (since labor is scarcer).

Goodhart–Pradhan's framing was a precocious forecast made just before the 2021 inflation surge. The post-2021 data are partially consistent: tight labor markets, persistent services inflation, and renewed wage growth all line up with the Goodhart–Pradhan picture.

But the Auclert et al. framework reaches a different conclusion: even with aging, the compositional effect (more wealthy savers) outweighs the labor-supply effect, so r-star stays low. The disagreement turns on subtleties of model specification — does aging raise the desired wealth-to-GDP ratio faster than it lowers labor productivity?

Bottom line. Demographics will matter a lot for the next 30+ years of macroeconomic dynamics; the sign and magnitude of the r-star and inflation effects are still being debated.

12. International macro: dollar dominance, capital flows, the trilemma

12.1 The dominant currency paradigm

Gita Gopinath's "Dominant Currency Paradigm" (AER 2020; review with Itskhoki 2022) reshaped international macroeconomics (Gopinath, Itskhoki review). The empirical fact: most non-US trade is invoiced in dollars, far in excess of the US share of world trade. This means:

• Exchange rate pass-through. When the dollar appreciates against everything, dollar-invoiced exports become more expensive for buyers and cheaper for sellers, even when neither party is American. Pass-through to import prices is fastest and largest for dollar-invoiced goods.
• Optimal monetary policy in emerging markets. Pegging to the dollar or managing dollar volatility may be welfare-improving when the dollar is the global invoice currency.
• The Triffin dilemma updated. Demand for dollar-denominated safe assets grows with global trade; US ability to supply them grows with US debt issuance. Excessive supply risks reserve-status erosion; insufficient supply risks global dollar funding squeezes (as in March 2020).

The IMF's 2025 patterns-of-invoicing paper confirms dollar dominance has not eroded materially despite years of de-dollarization rhetoric (IMF WP 2025/178).

12.2 The trilemma vs the dilemma

The classical Mundell–Fleming trilemma: a country can have any two of (i) fixed exchange rate, (ii) free capital mobility, (iii) independent monetary policy. Pick two.

Hélène Rey ("Dilemma not Trilemma," Jackson Hole 2013, NBER WP 21162) argued the trilemma has become a dilemma: under free capital mobility, monetary policy independence is severely constrained regardless of exchange rate regime, because the global financial cycle — driven primarily by US monetary policy — synchronizes capital flows, asset prices, and credit conditions across countries (Rey NBER WP 21162; CEPR VoxEU summary).

The 2024 Kim paper in Review of International Economics (Kim 2024) provides updated evidence supporting Rey's dilemma view, particularly for long-term interest rates: even floating-rate economies see their long yields move in lockstep with US yields.

Policy implication: emerging-market central banks need additional tools beyond the policy rate — FX intervention, macroprudential policy, capital flow management. The IMF's Integrated Policy Framework (IPF), formalized in 2020–2023, codifies this with explicit guidance on combining monetary policy, FX intervention, macroprudential measures, and capital flow management (IMF Integrated Policy Framework topic page).

12.3 Capital flow management

The IMF's 2012 Institutional View established capital flow management measures (CFMs) as legitimate policy tools in specific contexts. The 2022 review expanded the toolkit to allow pre-emptive use of CFMs on inflows when stock vulnerabilities threaten stability — a major intellectual reversal from the unrestricted-mobility orthodoxy of the 1990s (IMF "Review of Institutional View on the Liberalization and Management of Capital Flows" 2022).

A 2024 IMF working paper formalizes the framework for capital controls on outflows in crisis (IMF WP 2024/164).

12.4 BRICS+, de-dollarization, the geoeconomics turn

Geopolitical events 2022–2025 (Russia sanctions; US use of dollar-clearing as a sanctions tool; China's bilateral RMB-settlement push; BRICS expansion) have spurred discussion of de-dollarization. The empirical reality through 2026 is mostly: marginal share losses in international reserves (dollar reserves down from ~71% to ~58% of allocated reserves over 25 years) but dominant invoicing share remains intact. The biggest threats to dollar dominance are likely homegrown — US fiscal trajectory, threats to Fed independence, US export of dollar-clearing-as-weapon — rather than coordinated displacement by alternatives.

13. Climate macro

13.1 The DICE / IAM tradition

The Dynamic Integrated model of Climate and the Economy (DICE), developed by William Nordhaus since the 1990s, is the most-cited integrated assessment model (IAM). DICE couples a macro growth model (Ramsey-type) with a simple carbon cycle and a damage function mapping temperature to GDP losses.

DICE-2023 (Barrage and Nordhaus, NBER WP 31112 / PNAS 2024) updates the model with revised damage functions, lower-cost decarbonization assumptions, and lower discount rates. Key results (Barrage & Nordhaus NBER WP 31112):

• Optimal social cost of carbon (SCC): substantially higher than DICE-2016 estimates.
• A $80/tCO2 global carbon price is needed to meet the 2°C Paris target.
• Current global average is roughly $3/tCO2 — orders of magnitude too low.

DICE has been criticized for under-weighting tail risks (catastrophic outcomes), using inadequate damage functions for very high warming, and discount rate choices that may favor present over future generations (LSE Grantham Institute critique).

13.2 The Golosov–Hassler–Krusell–Tsyvinski (GHKT) sufficient-statistic formula

GHKT (Econometrica 2014) derive a remarkably clean closed-form expression for the optimal carbon tax in general equilibrium. Under specific (and not implausible) assumptions about the climate-damage relation being log-linear in atmospheric carbon and about preferences being logarithmic, the optimal carbon tax (per ton CO2) is:

τ = δ · GDP / (carbon depreciation parameter)

where δ captures the discount rate and the damage elasticity. Strikingly, future GDP, consumption, and CO2 paths do not enter — they cancel out. This is a "sufficient statistic" formula in the modern Saez–Chetty sense (GHKT Econometrica 2014).

Practical implication: the optimal carbon tax is approximately proportional to current GDP, scaling about 1:1 with economic growth. Doubling global GDP doubles the optimal carbon tax. Sensitivity analyses (Barrage 2014 supplement; many follow-ups) explore how this changes when assumptions are relaxed.

13.3 Directed technical change and the green transition

Acemoglu, Aghion, Bursztyn, and Hémous (AER 2012) and the extended Acemoglu–Aghion–Hémous North-South model show that with sufficient substitutability between "dirty" and "clean" inputs, sustainable growth requires only temporary, well-designed subsidies/taxes that redirect innovation toward clean technologies — eventually the clean sector dominates endogenously, and subsidies can be removed.

The 2024 Acemoglu, Aghion, Barrage, Hémous paper "Green Innovation and the Transition Toward a Clean Economy" updates this with quantitative analysis (SSRN 4816734). The takeaway: a credible carbon-price-plus-innovation-subsidy combination can deliver decarbonization at substantially lower cost than carbon-price-alone.

13.4 NGFS scenarios and central banks

The Network for Greening the Financial System (NGFS), now comprising 144 central banks and supervisors plus 21 observers, has become the standard provider of climate-macro scenarios for central bank stress testing.

Phase V scenarios (November 2024) project potential global GDP losses up to 30% by 2100 under current policies, with tail risks of up to 50% — more than double previous estimates. Several countries could see GDP shrink by a quarter as soon as 2050 from chronic climate risks alone (NGFS Phase V; Green Central Banking summary).

The NGFS short-term scenarios (May 2025) extend coverage through 2030 — the first publicly available dedicated framework for near-term climate-policy and climate-change financial stability analysis (NGFS short-term scenarios).

13.5 The "green QE" debate

Should central banks tilt their asset purchases toward green securities, accept green collateral on more favorable terms, or otherwise use balance-sheet tools to support decarbonization? Proponents (some ECB voices, French Finance Ministry, NGFS) argue climate risk is a financial-stability issue and so within mandate. Skeptics (Bundesbank, BoE under Bailey, much of the Fed) argue this risks compromising central bank independence by drawing it into industrial policy. The ECB under Lagarde has cautiously tilted its corporate bond purchases toward lower-carbon issuers; the Fed has explicitly rejected the practice.

14. Central bank independence under stress

Central bank independence has been the cornerstone of post-1990 inflation control. Kydland–Prescott's time-inconsistency analysis and Rogoff's "conservative central banker" provide the formal case: a central bank insulated from electoral incentives can credibly commit to disinflation in a way an electorally accountable Treasury cannot.

The 2025–2026 period is the most public stress test of independence in advanced economies in 50 years:

• United States. The Trump administration's pressure campaign on Powell — public threats to fire, demands for rate cuts, DOJ investigations, attempts to install loyalists on the Board — represents a sustained challenge unmatched in the postwar era (PBS Frontline coverage; Atlantic Council "Trump's challenges to the Fed's independence"). Blockchain prediction-market analysis suggests that when markets assign higher probability to Powell's dismissal, short-rate expectations fall (markets expect dovish replacement) but long-rate expectations rise (markets demand a term premium for less-credible inflation control) (CEPR VoxEU "Under pressure?").
• Joint statement of 10 central banks. In an unprecedented intervention, ten major central banks and global financial institutions issued a joint statement supporting Powell and the principle of central bank independence (CBS News coverage).
• ECB's independence is constitutionally entrenched and has not been similarly challenged, though Eurogroup-ECB tensions have surfaced periodically.
• BoE operates under a more politically embedded framework (government sets the target; BoE pursues it) and has faced regular parliamentary pressure but not systemic challenge.
• BoJ's independence has been operational rather than statutory, with significant historical political influence over governorships and policy.

The empirical literature (Cukierman, Garriga-Rodriguez, recent work using the CBI index) finds independence is correlated with lower and more stable inflation. The 2025–2026 episode is a test of whether the political-economy case for independence holds when challenged in real time.

15. Crypto, stablecoins, CBDC

15.1 Stablecoins as macro actors

Stablecoin market capitalization grew from <$2bn in 2019 to >$250bn by early 2025. Annual trading volume reached the trillions. Tether (USDT) and USDC dominate; both back tokens primarily with US Treasury bills (10,000-Word Analysis of Stablecoins).

Macro implications:

• Treasury market footprint. Tether and USDC together hold more US Treasuries than Saudi Arabia. Tether was the 7th largest net buyer of US Treasuries in 2024 (IMF 2025 ESR analysis). Research suggests every 1% increase in Tether's T-bill share lowers 1-month T-bill yields by ~3.8 basis points.
• Dollar-export channel. Stablecoins effectively export dollar-denominated money to emerging markets where local currencies are unstable. This is "digital dollarization" with no direct US oversight.
• Run risk. A run on a major stablecoin would force fire-sale of Treasury holdings, potentially disrupting short-end markets.

The US GENIUS Act (July 2025) imposed 100% reserves in cash or short-dated Treasuries for payment stablecoins, plus disclosure requirements — a major regulatory backstop (Elliptic stablecoin guide). The ECB's November 2025 FSR notes spillover risks even as euro-area direct exposure remains small (ECB FSR Nov 2025).

15.2 CBDC design tradeoffs

Central Bank Digital Currencies (CBDCs) are in production in several emerging markets (e.g., China's e-CNY, Nigeria's eNaira) and in advanced research in most advanced economies. Key design dimensions (arxiv survey 2507.08880; IMF policy paper 2025):

• Direct vs intermediated. Should households hold accounts directly at the central bank, or through commercial banks acting as intermediaries? Two-tier (intermediated) is the modal approach.
• Token-based vs account-based. Token-based offers privacy but creates verification challenges; account-based ties to identity.
• Interest-bearing vs non-interest-bearing. Interest-bearing CBDC is a more powerful monetary policy tool but more disruptive to commercial bank funding models.
• Wholesale vs retail. Wholesale CBDC (interbank only) is less controversial; retail CBDC raises bank-disintermediation and surveillance concerns.

The Fed has explicitly stated it will not issue a CBDC without congressional authorization. The ECB's digital euro project is in preparation phase, targeting potential launch in the late 2020s. The BoE has signaled openness to a "Britcoin" but with no fixed timeline.

Monetary policy implications: a sufficiently substitutable, possibly interest-bearing CBDC would constrain bank deposit pricing, potentially weaken the bank lending channel, but might strengthen the direct monetary transmission channel (Bank of Canada SWP 2024-27).

Mathematical foundations

This section presents the algebra. Where advanced math would be required to go further, I name the prerequisites and stop.

M.1 The Taylor rule

The standard Taylor (1993) rule:

i_t = r* + π* + φ_π · (π_t − π*) + φ_y · (y_t − y*)

Where:

• i_t is the nominal policy rate at time t
• r* is the equilibrium real rate
• π* is the inflation target (e.g., 0.02)
• π_t is current inflation
• y_t − y* is the output gap (deviation from potential)
• φ_π is the inflation reaction coefficient (Taylor proposed 1.5)
• φ_y is the output-gap reaction coefficient (Taylor proposed 0.5)

The Taylor principle: for determinacy, φ_π > 1. Otherwise the system has indeterminate equilibria — multiple inflation paths can satisfy the model.

Variants in current Fed practice include inertial Taylor rules (smoothing the change in i_t), unemployment-gap versions (replacing the output gap), and asymmetric rules (different φ_π for positive vs negative inflation deviations). BIS work shows that the Fed's effective 2022–2023 reaction function was asymmetric, reacting more aggressively to inflation overshoot than undershoot (BIS "Targeted Taylor rules" 2024).

M.2 The New Keynesian Phillips Curve (linearized)

The textbook NK Phillips curve (Galí 2008, Monetary Policy, Inflation, and the Business Cycle):

π_t = β · E_t[π_{t+1}] + κ · ỹ_t

Where:

• π_t is inflation
• E_t[π_{t+1}] is rational-expectations-conditional expectation of next-period inflation
• β is the household discount factor (close to 1)
• κ is the slope, derived from the deep parameters as κ = λ · (σ + (φ + α)/(1 − α)) where λ depends on the Calvo price-stickiness parameter, σ is the inverse elasticity of intertemporal substitution, φ is the inverse Frisch elasticity, α is the labor share complement
• ỹ_t is the output gap (relative to flexible-price natural output)

Intuition: today's inflation is anchored by expectations of next-period inflation plus a contribution from current excess demand. With well-anchored expectations (E_t[π_{t+1}] ≈ π*), inflation is close to target except for transient deviations driven by the output gap.

The HHNS "flat curve" finding is essentially that κ has been very small (~0.01) for decades. The Benigno–Eggertsson nonlinear finding adds a state-dependent component:

π_t = β · E_t[π_{t+1}] + κ(V/U) · ỹ_t

where κ(V/U) is small when V/U is low and large when V/U exceeds a threshold.

Study this next. Deriving the NK Phillips curve from optimizing-firm pricing under Calvo random-adjustment requires: (a) the firm's intertemporal pricing problem (a functional equation), (b) log-linearization around steady state, (c) aggregation across firms. The Galí textbook (3rd ed., 2015) is the standard derivation. To go deeper into HANK-style state-dependent pricing, see Auclert, Rognlie & Straub's work or Nakamura–Steinsson's review articles on price stickiness.

M.3 The IS curve from the Euler equation

The consumption Euler equation in a representative-agent NK model:

C_t^(−σ) = β · (1 + r_t) · E_t[C_{t+1}^(−σ)]

In words: at the optimum, the marginal utility of consumption today equals the discounted expected marginal utility tomorrow, scaled by the gross real interest rate.

Log-linearizing around the steady state and using market clearing (Y_t = C_t for simplicity), one obtains the dynamic IS curve:

ỹ_t = E_t[ỹ_{t+1}] − (1/σ) · (i_t − E_t[π_{t+1}] − r*)

The output gap depends on expected future output gap minus a term proportional to the real interest rate gap. Intuition: higher real rates depress current relative to future consumption.

Calculus warning. The Euler equation itself comes from taking the first-order condition of an intertemporal optimization with respect to consumption — using calculus. The log-linearization is a Taylor expansion. To work with the underlying optimization rigorously you need calculus and an understanding of dynamic programming (the Bellman equation form is V(state) = max [u(c) + β · E V(state′)]).

M.4 The fiscal theory identity

The FTPL government valuation identity:

B_t / P_t = E_t [ Σ_{k=0}^∞ ρ^k · s_{t+k} ]

Where:

• B_t is nominal government debt outstanding at end of t
• P_t is the price level
• ρ is the discount factor used to value future real primary surpluses
• s_{t+k} is the real primary surplus in period t+k (taxes minus non-interest spending)

The left side is real debt outstanding. The right side is the present value of real primary surpluses. The two must be equal in equilibrium.

The FTPL claim: if the right side falls (markets revise down expected surpluses), and if the central bank is not actively adjusting B_t to compensate, then P_t must rise. In other words, the price level adjusts to clear the government's intertemporal budget constraint.

In a Ricardian regime, the surpluses adjust to validate any debt-and-money trajectory the central bank chooses. In a non-Ricardian (fiscal-dominant) regime, the price level adjusts to validate the surpluses.

The empirical test is hard: surplus expectations are not directly observed. Cochrane (2023) and Bianchi–Faccini–Melosi (2023) approach this differently — Cochrane via the long-end yield curve and inflation expectations; BFM via a structural model identifying funded vs unfunded shocks.

M.5 Government debt accumulation

In discrete time, government debt evolves as:

B_t = (1 + i_{t-1}) · B_{t-1} − P_t · s_t

Dividing by P_t · Y_t (nominal GDP) and writing b_t = B_t / (P_t · Y_t):

b_t = ((1 + r_t) / (1 + g_t)) · b_{t-1} − s_t/Y_t

For small r and g this approximates to:

b_t ≈ b_{t-1} + (r − g) · b_{t-1} − pb_t

where pb_t = s_t/Y_t is the primary balance to GDP ratio. The change in the debt-to-GDP ratio depends on (i) the snowball term (r − g) · b_{t-1} — when r > g, debt grows; when r < g, debt shrinks even at balanced primary budget; (ii) the primary balance pb_t.

This is the math behind Blanchard's (r − g) analysis (§7.2). When (r − g) < 0, a country can run permanent primary deficits and still see debt/GDP fall.

M.6 The quantity theory — and why it isn't quite right anymore

Classical quantity theory:

M · V = P · Y    or, in growth rates,    m + v = π + y

Money growth (m) plus velocity growth (v) equals inflation (π) plus real output growth (y). With stable velocity, money growth determines nominal GDP growth.

The post-2008 problem: the Fed expanded its balance sheet from $900bn to $4.5tn (2008–2014) and then to $8.9tn (2020–2022). In a stable-velocity quantity theory, this should have produced massive inflation. It did not — because velocity collapsed (banks held reserves rather than lent; corporate cash hoards grew).

The modern critique (Reis, Curdia–Woodford 2011 and others): once central banks pay interest on reserves, the link from monetary base to broad money to nominal GDP breaks. The relevant transmission is via the policy rate and term-structure, not via base money. The "money multiplier" is not a structural relation but an artifact of regime.

This does not mean money doesn't matter — broad credit, lending, and the term structure all matter for the macroeconomy. It means the simple M · V = P · Y is misleading as a guide to inflation.

M.7 The HLW r-star intuition

The Holston–Laubach–Williams model treats r-star, potential output growth, and the output gap as latent (unobserved) variables in a state-space model. The observed data are: real GDP, inflation, the real federal funds rate.

The model assumes:

• The natural rate r*_t follows a random walk (perhaps with a deterministic trend tied to potential growth).
• The output gap ỹ_t is driven by an IS-like equation relating it to the real interest rate gap r_t − r*_t.
• A Phillips-curve-like relation links the output gap to inflation.

Given a state-space representation, the Kalman filter extracts the best estimate of r*_t at each date. The model is reestimated regularly; estimates are revised when new data arrive.

Study this next. State-space models and Kalman filtering: see Harvey (1989), Hamilton (1994 Time Series Analysis ch. 13), Durbin–Koopman (2012). The math requires linear algebra fluency, basic probability, and an understanding of recursive Bayesian updating. The reason the HLW estimates have wide error bands: r-star is identified only through model-based residual inference, and structural breaks (like COVID) can severely distort the filter.

M.8 The two-agent TANK in algebra

A simple version: a fraction λ of households are hand-to-mouth (H), spending all current income; a fraction (1−λ) are optimizing (O), spending according to standard permanent-income logic. Aggregate income equals aggregate consumption plus government spending (closed economy, no investment for simplicity):

Y = C_H + C_O + G

Hand-to-mouth consume:

C_H = Y_H

If government spending G is fully financed by lump-sum taxes T on optimizing households:

T_O = G, C_O ≈ Permanent Income · (1 − tax effect)

A balanced-budget fiscal expansion ΔG raises Y_H by Δ(Y_H) if Y_H rises proportionally with aggregate. With Y_H = (income share of H) · Y, and aggregate consumption identity, the multiplier becomes:

dY/dG = 1 / (1 − λ · (income elasticity of H))

If hand-to-mouth income rises one-for-one with aggregate income, and λ = 0.3 (US estimate), the simple multiplier is 1 / (1 − 0.3) = 1.43. If hand-to-mouth income is countercyclical (a recession hits them harder), the multiplier is larger.

This algebra captures the essential HANK insight that fiscal multipliers depend not just on hand-to-mouth shares but on how hand-to-mouth income covaries with aggregate income.

M.9 Sufficient statistics — the GHKT carbon tax intuition

GHKT's central result, stated informally:

optimal carbon tax (per ton CO2) = γ · current GDP

where γ collapses several deep parameters: the social discount rate, the marginal damage of atmospheric carbon, and the rate at which carbon depreciates from the atmosphere. The future paths of GDP, consumption, and emissions do not enter in the canonical case.

The mathematical reason: under log utility, the dependence of marginal damages on future GDP exactly cancels the dependence on future discounted consumption, leaving a clean ratio. This is an instance of "sufficient statistics" methodology in the Saez-Chetty sense: a few empirically estimable moments encode all the information needed for the optimal policy.

Study this next. The sufficient-statistic approach in macro and public economics is a major modern tool. References: Chetty (2009 Annual Review), Saez (various), and Hendren–Sprung-Keyser (2020 QJE on welfare-weighted MVPF analysis). The mathematical prerequisites are convex optimization and Lagrangian methods plus comparative statics in general equilibrium.

Risks, limitations, and open questions

What is genuinely contested

1. The slope and nonlinearity of the Phillips curve. Three live positions: "flat as ever" (HHNS), "nonlinear, steep when tight" (Benigno–Eggertsson, Cerrato–Gitti), "spurious nonlinearity from supply-shock mismeasurement" (NBER WP 33522). The choice has direct policy implications: the flat view supports activist make-up strategies; the nonlinear view warrants preemptive tightening before vacancies tighten too much.

2. The level of r-star. NY Fed (low) vs Richmond Fed (high) vs survey-based (high) vs market-implied (volatile but elevated). Disagreement is several percentage points. Given current policy rates around 4–5%, whether the stance is "restrictive" or "neutral" depends crucially on this number.

3. The fiscal share of post-2021 inflation. Cochrane / BFM: substantial fiscal contribution. Reis / Bernanke–Blanchard: minor fiscal contribution. The disagreement reflects deep theoretical differences (Ricardian vs non-Ricardian regime; how to identify "unfunded" shocks empirically).

4. AI's effect on productivity. Acemoglu's modest (~0.7% over a decade) vs more optimistic (several percent) forecasts. Will not be resolved without years more data.

5. The future direction of r-star from demographics. Auclert–Malmberg–Martenet–Rognlie (low and falling) vs Goodhart–Pradhan (rising as labor scarcity bites). Both make defensible cases.

6. The optimal balance-sheet target. What level of reserves counts as "ample"? The Fed's revealed answer (mid-late 2025) is around $2.7 trillion, well above pre-2008 levels, but the rationale is largely operational rather than welfare-theoretic.

7. Whether the dollar-dominance equilibrium is stable in the long run. Continuing share but eroding margins; geopolitical headwinds. Modeling the equilibrium is hard because it depends on coordination and history.

Genuine limitations of the frontier

• DSGE estimation is data-hungry and identification is fragile. Many "structural" parameters are weakly identified by the data alone; Bayesian priors do much of the work, and prior-sensitivity is high.

• HANK requires computational sophistication that limits reproducibility. The number of practitioners who can build, calibrate, and debug a full HANK model is small. Toolkits (Auclert et al.'s sequence-jacobian) help but don't fully democratize.

• Climate-macro IAMs make strong assumptions about climate science feedbacks. Damage functions are essentially calibrated by assumption rather than estimated. Tail risks are routinely under-weighted.

• r-star is fundamentally unobservable. Any policymaker using a single r-star number is making a strong assumption.

• Inflation expectations data are noisy. Household surveys (Michigan) tend to overestimate persistently; market-implied measures embed liquidity and risk premia; firm surveys (CBE) are newer and not yet stress-tested.

• The 2020–2024 macro experiment is a single observation. Generalizing from it is hazardous. Most lessons drawn from it have a "this time was different" alternative interpretation.

What is genuinely settled (or close to it)

• Inflation expectations matter enormously, and well-anchored expectations are precious. Anchoring did most of the work in the 2023–2024 disinflation.
• Independent central banks with credible inflation mandates produce lower and more stable inflation. The post-2025 stress test of this is ongoing.
• Fiscal multipliers are positive and meaningfully above zero, especially in slack economies with constrained monetary policy.
• Heterogeneity matters for monetary and fiscal transmission. The RANK approximation, while still useful pedagogically, fails empirically in important ways.
• The trilemma is more like a dilemma; the global financial cycle is real.
• Demographic forces are large, slow-moving, and affect r-star and growth materially.
• Climate change has measurable, growing macroeconomic costs that justify substantial carbon pricing.

Most important open questions for the next 5 years

• How durable is the labor market's apparent willingness to absorb cooling without a sharp rise in unemployment? (Beveridge curve dynamics tested by ongoing slowdown.)
• Will the Fed's 2025 framework hold up if inflation drifts back above 2.5%? Or will it be revised again?
• Is the global financial cycle weakening (as some 2024–2025 evidence suggests) or just temporarily disrupted by the post-COVID monetary shock?
• Can central bank independence survive the current US political stress test?
• Will AI-driven productivity show up in measured TFP within the next 5 years, or will the J-curve dynamic delay it?
• Will the demographic transition raise inflation or just compress growth (or both)?
• What happens to global Treasury demand if stablecoins continue to grow at current rates and become major holders?

Recommendations for further study

A staged reading path organized by difficulty.

Stage 1: foundations (textbook level)

• Galí, Monetary Policy, Inflation, and the Business Cycle (3rd ed., 2015). The canonical NK textbook. Requires comfort with calculus and basic dynamic optimization. Difficulty: medium.
• Romer, Advanced Macroeconomics (5th ed., 2019). Broader graduate-text coverage including growth, RBC, NK, fiscal. Difficulty: medium.
• Blanchard, Macroeconomics (8th ed., 2021). Undergraduate-level but currently the best policy-oriented introduction. Difficulty: low.
• Krugman & Wells, Macroeconomics. Even more accessible, useful for intuition-building. Difficulty: low.

Stage 2: research-frontier overviews (advanced undergraduate / early graduate)

• Bernanke, 21st Century Monetary Policy: The Federal Reserve from the Great Inflation to COVID-19 (Norton, 2022). Best single-volume policy-history-and-mechanics. Difficulty: low-medium.
• Blanchard, Fiscal Policy under Low Interest Rates (MIT Press, 2023). The state-of-the-art on fiscal sustainability. Difficulty: medium.
• Reis, "How do central banks control inflation? A guide for the perplexed" (LSE/CFM, 2023). Synthesizes the alternative theoretical frameworks. Difficulty: medium-high.
• Ramey, "Macroeconomic Shocks and Their Propagation" (Handbook of Macro, 2016). Comprehensive survey of identification methods. Difficulty: high.

Stage 3: HANK and heterogeneity

• Kaplan & Violante, "The Marginal Propensity to Consume in Heterogeneous Agent Models" (Annual Review of Economics, 2022). Best entry-level review. Difficulty: medium-high.
• Kaplan, Moll & Violante, "Monetary Policy According to HANK" (AER 2018). The foundational HANK paper. Difficulty: high.
• Auclert, Bardóczy, Rognlie & Straub, "Using the Sequence-Space Jacobian" (Econometrica 2021). The computational tooling. Difficulty: very high.
• Auclert, Rognlie & Straub, "The Intertemporal Keynesian Cross" (JPE 2024). The current iMPC framework. Difficulty: high.

Stage 4: fiscal theory

• Cochrane, The Fiscal Theory of the Price Level (Princeton, 2023). The book-length statement. 584pp. Difficulty: medium-high (clearly written but mathematically dense in parts).
• Bianchi, Faccini & Melosi, "A Fiscal Theory of Persistent Inflation" (QJE 2023). Recent quantitative application. Difficulty: high.

Stage 5: climate macro

• Nordhaus & Barrage, "Policies, Projections, and the Social Cost of Carbon: Results from the DICE-2023 Model" (PNAS 2024). Current state of integrated assessment. Difficulty: medium.
• Golosov, Hassler, Krusell & Tsyvinski, "Optimal Taxes on Fossil Fuel in General Equilibrium" (Econometrica 2014). Sufficient-statistic carbon tax. Difficulty: high.
• NGFS Phase V Scenarios (November 2024) and Short-Term Scenarios (May 2025). Operational central bank scenarios. Difficulty: low-medium.

Stage 6: ongoing reading

• Fed FEDS Notes (federalreserve.gov/econres/notes/feds-notes). Short, policy-oriented working papers.
• Brookings Papers on Economic Activity (brookings.edu/papers-on-economic-activity). Two issues per year of state-of-the-art applied macro.
• Jackson Hole Symposium papers (kansascityfed.org/research/jackson-hole-economic-symposium). Annual showcase of frontier policy debates.
• NBER Working Papers (Macro programs). High volume; filter aggressively.
• Liberty Street Economics (NY Fed blog). Excellent for plumbing and balance-sheet topics.
• Conversable Economist (Tim Taylor) and John Cochrane's "Grumpy Economist" for editorial perspective.
• Apricitas (Joseph Politano) and Employ America (Skanda Amarnath et al.) for high-quality data analysis.

Mathematical prerequisites for serious participation

If your goal is to read and verify frontier papers, you need:

• Real analysis (single variable: Rudin; multivariate: Munkres).
• Linear algebra at the level of Strang or Axler.
• Probability and measure theory at the level of Billingsley or Williams.
• Dynamic programming (Stokey–Lucas–Prescott Recursive Methods; Adda–Cooper).
• Time series (Hamilton).
• Bayesian econometrics (Koop, Bayesian Econometrics).
• Numerical methods (Judd, Numerical Methods in Economics).

For HANK specifically, add: functional analysis basics, sequence-space methods (Auclert et al.'s online appendices and the sequence-jacobian Python toolkit), and PDE intuition (since continuous-time HANK uses HJB equations).

Source inventory

Frameworks (DSGE, HANK, TANK, FTPL, RBC)

• Smets & Wouters, "Shocks and Frictions in US Business Cycles: A Bayesian DSGE Approach," AER 2007, 97(3): 586–606.
• Kaplan, Moll & Violante, "Monetary Policy According to HANK," AER 2018, 108(3): 697–743.
• Auclert, Bardóczy, Rognlie & Straub, "Using the Sequence-Space Jacobian to Solve and Estimate Heterogeneous-Agent Models," Econometrica 2021, 89(5): 2375–2408.
• McKay, Nakamura & Steinsson, "The Power of Forward Guidance Revisited," AER 2016, 106(10): 3133–3158.
• Auclert, Rognlie & Straub, "The Intertemporal Keynesian Cross," JPE 2024, 132(12).
• Bilbiie, "The New Keynesian Cross," JME 2020.
• Cochrane, The Fiscal Theory of the Price Level, Princeton University Press, 2023.
• Bianchi, Faccini & Melosi, "A Fiscal Theory of Persistent Inflation," QJE 2023, 138(4): 2127–2179.
• Reis & Castillo-Martinez, "How Do Central Banks Control Inflation? A Guide for the Perplexed," LSE/CFM, 2023.
• Mankiw & Reis, "Sticky Information versus Sticky Prices," QJE 2002, 117(4): 1295–1328.

Post-2021 inflation

• Bernanke & Blanchard, "What Caused the U.S. Pandemic-Era Inflation?" Brookings Hutchins WP, 2023.
• Bernanke & Blanchard, "An Analysis of Pandemic-Era Inflation in 11 Economies," Brookings WP91, May 2024.
• Shapiro, "Decomposing Supply and Demand Driven Inflation," FRBSF WP 2022-18; data tool.
• Shapiro, "Is Demand or Supply More Important for Inflation?" FRBSF Economic Letter, June 2025.
• Reis, "The Burst of High Inflation in 2021–22: How and Why Did We Get Here?" LSE, 2022.
• Weber & Wasner, "Sellers' Inflation, Profits and Conflict: Why Can Large Firms Hike Prices in an Emergency?" Review of Keynesian Economics 2023.
• Konczal & Lusiani, "Prices, Profits, and Power: An Analysis of 2021 Firm-Level Markups," Roosevelt Institute, June 2022.
• Alvarez et al., "Wage-Price Spirals: What is the Historical Evidence?" IMF WP 2022/221.
• Lorenzoni & Werning, "Wage Price Spirals," IMF/AME seminar paper, August 2023.
• Stiglitz & Regmi, "The Causes of and Responses to Today's Inflation," Roosevelt Institute, December 2022.
• Goolsbee, "Lessons from the Supply Side," Chicago Fed speech, February 5, 2025.
• Council of Economic Advisers, "Chain Reaction: 'Immaculate' Disinflation and the Role of Easing Supply Chains," Biden CEA, August 2023.
• Coibion & Gorodnichenko, "Inflation, Expectations and Monetary Policy: What Have We Learned and to What End?" Berkeley working paper; IZA DP 17919, 2025.
• Candia, Coibion & Gorodnichenko, "The Inflation Expectations of U.S. Firms," NBER WP 28836.

Phillips curve and Beveridge curve

• Hazell, Herreño, Nakamura & Steinsson, "The Slope of the Phillips Curve: Evidence from U.S. States," QJE 2022, 137(3): 1299–1344.
• Benigno & Eggertsson, "It's Baaack: The Surge in Inflation in the 2020s and the Return of the Non-Linear Phillips Curve," NBER WP 31197, revised December 2024.
• Eggertsson & Benigno, "Revisiting the Phillips and Beveridge Curves: Insights from the 2020s," NBER WP 33095 / Jackson Hole 2024.
• Cerrato & Gitti, "Nonlinearities in the Regional Phillips Curve with Labor Market Tightness," Cleveland Fed, 2024.
• Harding, Lindé & Trabandt, "On the Fragility of the Nonlinear Phillips Curve Interpretation of Recent Inflation," NBER WP 33522, 2025.
• Blanchard, Domash & Summers, "Bad News for the Fed from the Beveridge Space," PIIE, July 2022.
• Domash & Summers, "How Tight are U.S. Labor Markets?" NBER WP 29739.
• Figura & Waller, "What Does the Beveridge Curve Tell Us about the Likelihood of a Soft Landing?" FEDS Notes, July 29, 2022.

Monetary policy frameworks and tools

• Powell, "Monetary Policy and the Economic Outlook," Jackson Hole speech, August 22, 2025.
• Federal Reserve Board, "Review of Monetary Policy Strategy, Tools, and Communications" 2025 page.
• FEDS Note, "A Roadmap for the Federal Reserve's 2025 Review," August 22, 2025.
• ECB, "ECB's Governing Council Updates Its Monetary Policy Strategy," June 30, 2025 press release.
• ECB, "An Overview of the ECB's Monetary Policy Strategy" June 2025.
• ECB, "ECB's Governing Council Approves Its New Monetary Policy Strategy," July 8, 2021 press release.
• Bank of Japan, "Changes in the Monetary Policy Framework," March 19, 2024 statement.
• Reifschneider & Williams, "Three Lessons for Monetary Policy in a Low Inflation Era," Boston Fed conference 2000.

Balance sheet, plumbing, QE/QT

• NY Fed Teller Window, "Standing Repo Operations in the Federal Reserve's Monetary Policy Implementation Framework," December 23, 2025.
• Cleveland Fed, "QT, Ample Reserves, and the Changing Fed Balance Sheet," Economic Commentary 2025-05.
• Federal Reserve, "Bank Deposit Flows to Money Market Funds and ON RRP Usage during Monetary Policy Tightening," FEDS Note.
• PIIE, "How much money have central banks really lost?" 2025.
• CETEx, "The Bank of England's Asset Purchase Facility," September 2024.
• Acharya et al., "Banking Turmoil and Regulatory Reform," CEPR/IESE Future of Banking, 2024.

r-star, secular stagnation, ELB

• New York Fed, "Measuring the Natural Rate of Interest," HLW model page.
• Williams, "R-Star: A Global Perspective," July 3, 2024 speech.
• Richmond Fed, "Examining the Differences in r* Estimates," Economic Brief 24-36.
• Liberty Street Economics, "The Post-Pandemic Global R*," February 2026.
• Eggertsson, Mehrotra & Robbins, "A Model of Secular Stagnation: Theory and Quantitative Evaluation," AEJ:Macro 2019, 11(1): 1–48.
• Summers–Blanchard debate, PIIE transcript, March 2023.
• Blanchard, Fiscal Policy under Low Interest Rates, MIT Press, January 2023.

Fiscal policy

• Ramey, "Ten Years After the Financial Crisis: What Have We Learned from the Renaissance in Fiscal Research?" JEP 2019, 33(2): 89–114.
• Chodorow-Reich, "Geographic Cross-Sectional Fiscal Spending Multipliers: What Have We Learned?" AEJ:Policy 2019, 11(2): 1–34.
• Blanchard, "Public Debt and Low Interest Rates," AER 2019, 109(4): 1197–1229.
• Mankiw, "A Skeptic's Guide to Modern Monetary Theory," AEA P&P 2020.

Heterogeneity, MPCs

• Mian, Straub & Sufi, "The Saving Glut of the Rich," NBER WP 26941.
• Sufi, "What Explains the Decline in r*? Rising Income Inequality versus Demographic Shifts," Jackson Hole 2022.

Climate macro

• Barrage & Nordhaus, "Policies, Projections, and the Social Cost of Carbon: Results from the DICE-2023 Model," PNAS 2024 / NBER WP 31112.
• Golosov, Hassler, Krusell & Tsyvinski, "Optimal Taxes on Fossil Fuel in General Equilibrium," Econometrica 2014, 82(1): 41–88.
• Acemoglu, Aghion, Bursztyn & Hémous, "The Environment and Directed Technical Change," AER 2012, 102(1): 131–166.
• Acemoglu, Aghion, Barrage & Hémous, "Green Innovation and the Transition Toward a Clean Economy," SSRN 4816734.
• NGFS, "NGFS Climate Scenarios for Central Banks and Supervisors — Phase V," November 2024.
• NGFS, "Short-Term Climate Scenarios," May 2025.

Demographics

• Auclert, Malmberg, Martenet & Rognlie, "Demographics, Wealth, and Global Imbalances in the Twenty-First Century," NBER WP 29161; Jackson Hole 2024 "Race Between Asset Supply and Asset Demand".
• Goodhart & Pradhan, The Great Demographic Reversal: Ageing Societies, Waning Inequality, and an Inflation Revival (Palgrave Macmillan, 2020).

International macro

• Gopinath & Itskhoki, "Dominant Currency Paradigm: A Review," 2022 handbook chapter.
• Gopinath, Boz, Casas, Díez, Gourinchas & Plagborg-Møller, "Dominant Currency Paradigm," AER 2020.
• Rey, "Dilemma not Trilemma: The Global Financial Cycle and Monetary Policy Independence," Jackson Hole 2013 / NBER WP 21162.
• IMF, "Integrated Policy Framework" topic page.
• IMF, "Review of the Institutional View on the Liberalization and Management of Capital Flows," March 2022.

Productivity

• Brynjolfsson, Rock & Syverson, "The Productivity J-Curve: How Intangibles Complement General Purpose Technologies," AEJ:Macro 2021.
• Acemoglu, Autor & Patterson, "Bottlenecks: Sectoral Imbalances and the US Productivity Slowdown," NBER WP 31427.
• Yagan, "Employment Hysteresis from the Great Recession," JPE 2019.

Crypto, stablecoins, CBDC

• IMF, "Stablecoins, Tokens, and Global Dominance," F&D September 2025.
• ECB, "Stablecoins on the Rise: Still Small in the Euro Area, but Spillover Risks Loom," FSR Box, November 2025.
• "Central Bank Digital Currencies: A Survey," arXiv 2507.08880.
• IMF, "Central Bank Digital Currency: Further Considerations," Policy Paper 2025.

Central bank independence

• CEPR VoxEU, "Under Pressure? Central Bank Independence Meets Blockchain Prediction Markets," 2025.
• Atlantic Council, "Trump's Challenges to the Fed's Independence Loom over Jackson Hole Symposium," August 2025.
• PBS Frontline, "Inside Trump's Pressure Campaign on Powell and the Fed," 2025.

BIS overview pieces

• BIS, "So Far, So Good… 94th Annual Economic Report 2024," overview.
• BIS, "Monetary Policy in the 21st Century: Lessons Learned and Challenges Ahead," Chapter II of AER 2024.
• BIS, "Targeted Taylor Rules: Monetary Policy Responses to Demand- and Supply-Driven Inflation," Quarterly Review, December 2024.

Glossary appendix

Plain-English definitions of 100+ key terms. Cross-references in bold.

• Ample reserves regime. A monetary-policy operating framework in which the central bank holds enough reserves that overnight rates are set by administered rates (interest on reserves, RRP), not by reserve scarcity. The Fed's current framework.
• Asset Purchase Facility (APF). The BoE's QE vehicle; operates with a Treasury indemnity that has triggered large fiscal transfers since 2023.
• Average Inflation Targeting (AIT). A make-up strategy in which the central bank tolerates above-target inflation following below-target periods to keep the average at target. The Fed adopted FAIT in 2020 and abandoned it in 2025.
• Backstop facility. A central bank facility (like the Fed's SRF/SRP) that lets eligible counterparties borrow at a known rate, capping market rate spikes.
• Balance sheet policy. Central bank actions affecting the size and composition of its balance sheet (QE, QT, RRP, repo operations) distinct from setting the policy rate.
• Bayesian estimation. A statistical approach that combines prior beliefs with observed data; standard in modern DSGE estimation.
• Beveridge curve. The downward-sloping relation between unemployment (U) and vacancies (V). Shifted outward dramatically after COVID; central to the soft-landing debate.
• BIS (Bank for International Settlements). The "central bank for central banks" based in Basel; publishes influential macro research.
• BRICS+. The expanded grouping of Brazil, Russia, India, China, South Africa, plus Saudi Arabia, UAE, Egypt, Ethiopia, Iran since 2024. Periodically discussed as a potential de-dollarization vehicle.
• CBDC (Central Bank Digital Currency). A digital form of central bank money for retail or wholesale use. Operational in some EMs; in research/pilot in most advanced economies.
• CET (Common Equity Tier 1). Highest-quality bank capital under Basel rules.
• Cochrane fiscal theory. The view that the price level adjusts to satisfy the government's intertemporal budget constraint; central to modern FTPL revival.
• Coibion–Gorodnichenko expectations. Empirical body of work using surveys to document households' and firms' inflation expectations and their effects on decisions.
• Conservative central banker (Rogoff). Theoretical argument that delegating monetary policy to an inflation-averse central banker improves welfare by solving time-inconsistency.
• Convexity (Phillips curve). The state-dependent nonlinearity where the curve steepens as labor markets tighten.
• DSGE (Dynamic Stochastic General Equilibrium). The dominant quantitative framework in macro since the 1990s. Models a complete economy with optimizing agents, shocks, and frictions.
• Dual mandate. The Fed's statutory objectives: maximum employment and stable prices.
• ECB (European Central Bank). Central bank for the euro area. Confirmed 2% symmetric inflation target in 2021; reconfirmed in 2025.
• Effective Lower Bound (ELB). The lowest practical level for the policy rate, slightly above the strict ZLB because of cash-hoarding and operational frictions.
• Euler equation. The intertemporal optimization condition relating today's marginal utility of consumption to discounted expected future marginal utility, scaled by the real rate.
• FAIT (Flexible Average Inflation Targeting). Fed's 2020–2025 framework. Junked at Jackson Hole 2025.
• FedNow. The Fed's instant payment system (live since 2023), distinct from CBDC.
• FIT (Flexible Inflation Targeting). The Fed's restored framework as of August 2025.
• Forward guidance. Communication about the future path of policy rates intended to influence current behavior.
• Forward guidance puzzle. The standard NK model's prediction that distant future rate commitments have implausibly large current effects. Partially resolved in HANK.
• FOMC (Federal Open Market Committee). The Fed's monetary policy committee.
• FRB/US. The Fed's semi-structural model used alongside DSGEs.
• FRBSF (Federal Reserve Bank of San Francisco). Notable for Shapiro's inflation decomposition.
• FTPL (Fiscal Theory of the Price Level). The view that the price level is determined by the fiscal-monetary equilibrium identity (debt/PV-of-surpluses).
• General-purpose technology (GPT). A technology with broad applicability across the economy (electricity, IT, AI). Central to productivity J-curve dynamics.
• GHKT formula. Golosov–Hassler–Krusell–Tsyvinski's sufficient-statistic optimal carbon-tax formula.
• Greedflation. Popularized term for inflation attributed to corporate markup expansion.
• HANK (Heterogeneous Agent New Keynesian). NK models with explicit household heterogeneity in income, wealth, and constraints.
• Hand-to-mouth household. Household that spends all current income, typically due to low liquid wealth.
• Hazell–Herreño–Nakamura–Steinsson (HHNS). QJE 2022 paper finding the Phillips curve has been very flat for decades.
• HICP (Harmonised Index of Consumer Prices). The ECB's headline inflation measure.
• HLW (Holston–Laubach–Williams). State-space r-star model maintained by the NY Fed.
• Hysteresis. Permanent supply-side effects of demand-driven recessions.
• Immaculate disinflation. Phrase describing the unexpected disinflation of 2023–2024 without a US recession.
• Inflation expectations. Beliefs about future inflation. Anchoring them is widely viewed as the central job of credible central banks.
• Inflation targeting. Policy framework with an explicit numerical inflation goal; standard in advanced economies since 1990s.
• Integrated Policy Framework (IPF). IMF framework combining monetary, FX, macroprudential, and capital flow tools.
• Interest on Reserve Balances (IORB). The rate the Fed pays banks on reserves. Effective floor for the federal funds rate.
• Intertemporal MPC (iMPC). The full impulse response of consumption to a transitory income shock; sufficient statistic for fiscal multipliers in HANK.
• JGB (Japanese Government Bond).
• JOLTS (Job Openings and Labor Turnover Survey). US data source on vacancies, quits, hires.
• Kalman filter. Recursive Bayesian algorithm for extracting unobserved variables from time series; basis of HLW r-star estimation.
• Kydland–Prescott. Real Business Cycle pioneers; also formalized the time-inconsistency problem in monetary policy.
• Liquidity trap. Situation where conventional monetary policy is ineffective because rates cannot fall further. Functionally synonymous with binding ELB.
• Macroprudential policy. Regulatory tools (capital, liquidity, LTV/DTI limits) used to address financial-stability risks.
• Marginal propensity to consume (MPC). Fraction of an income increment spent (rather than saved). Higher for hand-to-mouth households.
• MMT (Modern Monetary Theory). Heterodox view emphasizing monetary sovereignty's role in fiscal capacity.
• NAIRU. Non-Accelerating Inflation Rate of Unemployment. Theoretical level of unemployment consistent with stable inflation.
• NAWM-II. The ECB's New Area-Wide Model, current vintage; the ECB's main DSGE.
• New Keynesian (NK). The post-1990s synthesis combining real-business-cycle optimization with nominal rigidities.
• NGFS (Network for Greening the Financial System). Multi-central-bank network producing standardized climate scenarios.
• Non-Ricardian regime. A fiscal-monetary configuration where the fiscal authority does not adjust future surpluses to validate the price level; the regime that activates FTPL.
• ON RRP (Overnight Reverse Repo). Fed facility letting non-banks invest cash overnight at administered rate. Provides a floor for non-bank money market rates.
• Open market operations (OMOs). Central bank purchases/sales of securities to manage reserves and rates.
• OLG (Overlapping Generations). Modeling framework with multiple co-existing age cohorts. Standard for demographic macro analysis.
• PCE (Personal Consumption Expenditures). The Fed's preferred inflation measure (vs CPI).
• Phillips curve. Inflation-unemployment trade-off. Currently understood as flat at slack and steep at tightness.
• Policy normalization. Process of moving policy rates from emergency lows back to neutral.
• Potential output. Maximum sustainable output consistent with stable inflation.
• Primary balance. Government budget balance excluding interest payments.
• QE (Quantitative Easing). Central bank asset purchases beyond normal OMOs, typically targeting longer-dated securities.
• QT (Quantitative Tightening). Reduction of central bank balance sheet, usually by letting securities roll off rather than reinvesting.
• Quantity theory of money (M·V = P·Y). Classical relation; modern critique notes velocity instability under interest-on-reserves regime.
• RANK (Representative Agent NK). Standard NK with one representative household. Pre-HANK workhorse.
• Rational expectations. Forward-looking expectations consistent with the model's predictions.
• Rational inattention. Theory (Sims) that agents face costs of information processing, leading to bounded responsiveness.
• Real Business Cycle (RBC). Pre-NK framework attributing business cycles to productivity shocks in frictionless economies.
• Real interest rate. Nominal rate minus expected inflation. The variable relevant for intertemporal substitution.
• Reserves (bank reserves). Bank deposits at the central bank. Central to the operating framework.
• Reverse repo (RRP). Sale of securities with agreement to repurchase, used by Fed as floor instrument.
• R-star (r).* Equilibrium real interest rate consistent with output at potential and stable inflation.
• Sequence-space Jacobian. Auclert et al. computational method for solving HANK models in the time domain.
• Sellers' inflation. Weber-Wasner term for inflation driven by firms exploiting market power amid supply shocks.
• Shapiro decomposition. Identifies inflation contributions from supply vs demand factors using price-quantity comovement.
• Smets–Wouters model. Canonical estimated NK DSGE; reference benchmark at major central banks.
• Soft landing. Disinflation without recession. Achieved (or nearly so) in 2023–2024 in the US.
• Standing Repo Facility (SRF) / Standing Repo Program (SRP). Fed backstop for short-term funding markets. Renamed and expanded December 2025.
• Sticky information. Mankiw-Reis model in which agents update their information set infrequently rather than prices being sticky.
• Sticky prices (Calvo). NK assumption that only a fraction of firms can reset prices each period.
• Stockholm school. Mid-20th-century macro tradition particularly strong on monetary theory.
• Sufficient statistic approach. Methodology in which a few empirically estimable moments encode the information needed for policy analysis.
• Supply chain pressures. Disruptions in international trade logistics; major driver of 2021–2022 inflation.
• Taper tantrum. 2013 episode of bond yield spike triggered by Fed signals of QE reduction.
• TANK (Two-Agent NK). Stripped-down HANK with two household types (hand-to-mouth and optimizers).
• Taylor rule. Reaction function describing how policy rates respond to inflation and output gaps.
• Taylor principle. Requirement that the policy rate respond more than one-for-one to inflation for determinacy.
• Term premium. Compensation investors require for holding longer-duration debt beyond pure expectations.
• Time inconsistency. Tension between optimal ex-ante and ex-post policy; central case for rules and independence.
• Trilemma (Mundell-Fleming). Country can have any two of: fixed exchange rate, free capital flows, monetary autonomy.
• Treasury indemnity (BoE). The UK Treasury's commitment to absorb gains/losses from the BoE's Asset Purchase Facility.
• UMich expectations. University of Michigan Survey of Consumers measure of household inflation expectations.
• Unfunded fiscal shock. Bianchi-Faccini-Melosi: expenditure shock with no expected future fiscal adjustment; activates fiscal-channel inflation.
• Vacancy-unemployment (V/U) ratio. Measure of labor market tightness used in modern Phillips curve analysis.
• Volcker disinflation. 1979–1982 Fed-led disinflation under Chair Paul Volcker; foundational case study.
• Wage-price spiral. Self-reinforcing dynamic of wage gains feeding price gains feeding wage gains. Largely contained in 2022–2024 episode.
• Wealthy hand-to-mouth. Households with high illiquid wealth but low liquid wealth; high MPCs despite high net worth.
• Yield curve. Term structure of interest rates on government debt.
• Yield Curve Control (YCC). BoJ policy 2016–2024 of targeting specific yields along the curve. Exited March 2024.
• Zero Lower Bound (ZLB). Theoretical floor at zero for nominal rates. Functionally synonymous with ELB for practical purposes.

End of report.

Cross-references to parallel reports in this series:

• History of macroeconomics (01-macroeconomics-history.md): for Keynes, Friedman, Lucas, Volcker, rational expectations, and the post-WWII evolution of mainstream macro that forms the prerequisite reading.
• State-of-the-art microeconomics (03-microeconomics-sota.md): for the heterogeneity-related micro-evidence (lottery studies, MPC estimation, search-and-matching frictions) that underpins HANK and TANK.
• Micro–macro intermingling (04-micro-macro.md): for deeper treatment of how distributional and firm-level evidence is being absorbed into macro models.
• Trading and applied macroeconomics (05-trading-economics.md): for the implications of these debates for asset prices, interest rate strategy, and the term structure.
• Master glossary (06-glossary.md): consolidated cross-report definitions.