Chapter 1: The Scalar Economy

The Shape of the Problem

Something has gone wrong with how we measure economic life — not the substance, but the form.

For most of its history, economics has operated with a tacit assumption: that economic evaluation, at the moment of decision, reduces to a single number. Utility theory makes this explicit: assign a real-valued utility to each outcome, choose the outcome with the highest utility. Cost-benefit analysis makes it operational: monetize all costs and benefits, sum them, and compare the totals. GDP makes it national: add up all market transactions in an economy and call the sum “economic output.”

The output is always a number. A ranking. A line.

This book argues that the assumption is wrong. Not because economic decisions are vague or irrational or too complex to model, but because they have geometric structure that a single number cannot represent. Economic evaluation is not a point on a line. It is a location in a space — a space with dimensions, distances, directions, curvature, and boundaries. When we flatten this structure into a scalar, we lose information. And the information we lose is precisely the information that matters most in the decisions that matter most: which values are at stake, where the moral boundaries lie, how trust and fairness interact with price, and why the same monetary gamble produces different choices depending on how it is described.

The mathematical name for this structure is geometry. Not the geometry of triangles and circles, but the far richer geometry of manifolds, tensors, metrics, and fiber bundles — the mathematics that describes how matter curves spacetime, how electromagnetic fields transform under change of coordinates, and how conservation laws emerge from symmetry. Geometric Economics argues that economic reality has exactly this character.

Three Failures of the Scalar

The limitations of scalar economic evaluation are not merely theoretical. They manifest in three domains where the stakes are highest.

The Easterlin Paradox and Its Geometric Resolution

Before turning to what geometry provides, consider a paradox that crystallizes the scalar failure.

In 1974, Richard Easterlin documented a striking finding: beyond a threshold of approximately $15,000 per capita (in 2024 dollars, roughly $75,000), national increases in GDP do not predict increases in reported life satisfaction. Japan’s real GDP per capita quadrupled between 1960 and 1990; average life satisfaction did not budge. The United States grew richer than any large nation in history; its happiness surveys flatlined from the 1970s onward, even as GDP per capita tripled.

[Empirical.] The standard explanations — hedonic adaptation, relative income effects, aspiration adjustment — describe the phenomenon without explaining it structurally. Why should there be a threshold? Why should money stop working at a particular level? Why should the relationship between income and happiness be logarithmic rather than linear?

The geometric framework provides a structural answer. Below the threshold, monetary growth (d_1) is strongly correlated with improvements on other dimensions: security (d_9), autonomy (d_4), social participation (d_6), basic rights (d_2). The covariance matrix \Sigma has large positive off-diagonal entries between d_1 and these other dimensions. Increasing d_1 pulls the entire attribute vector upward. GDP growth moves the economy along a geodesic that improves most dimensions simultaneously.

Above the threshold, the correlations weaken. Additional income does not automatically improve trust (d_5), fairness (d_3), community cohesion (d_6), or identity (d_7). The off-diagonal entries between d_1 and d_5–d_9 approach zero — or even become negative, as when economic growth undermines community bonds or intensifies status competition. GDP growth continues to move the economy along d_1, but the other eight dimensions are no longer carried along. The scalar says the economy is growing. The manifold says it is moving sideways.

The Easterlin paradox is not a paradox. It is the expected behavior of a system where the correlation structure between dimensions changes as a function of position on the manifold. Below the threshold, scalar optimization happens to be a reasonable proxy for full-manifold optimization because the dimensions are coupled. Above the threshold, the proxy fails because the coupling breaks down. GDP continues to be measured accurately — the measurement is not wrong. But it is measuring the wrong thing: a projection that has decoupled from the object it was meant to represent.

MARIA’S COFFEE SHOP — THE EASTERLIN PARADOX IN MINIATURE

Maria’s coffee shop nets $4,200 per month after the old rent. If her net income fell to $1,500, everything would degrade — she would cut employee hours (d_4), stop buying fair-trade beans (d_3, d_7), lose her community role (d_6), and live in constant financial anxiety (d_9). At $1,500, raising income to $4,200 would improve all nine dimensions. The correlation between d_1 and d_5–d_9 is strong.

But if Maria’s net income rose from $4,200 to $8,000 — if she doubled her revenue by converting to a franchise model — the additional money might come at the cost of autonomy (d_4: franchise rules replace her judgment), identity (d_7: she is no longer "Maria's place" but "Starbucks #4287"), community (d_6: the regulars leave when the character changes), and trust (d_5: her employees feel betrayed). More money, less welfare on the full manifold. The Easterlin paradox, at the scale of one small business.

The Monetization Trap

The three scalar failures — GDP, utility maximization, and cost-benefit analysis — share a common mechanism that deserves its own name. We call it the monetization trap: the assumption that all relevant information about an economic state can be expressed in a common monetary unit.

The monetization trap is seductive because it is sometimes correct. When the decision involves only d_1 — when the other dimensions are inactive or approximately invariant — monetization is lossless. Choosing between two savings accounts with different interest rates is a pure d_1 decision. Comparing the price of gasoline at two adjacent stations is a pure d_1 decision. In these cases, the scalar projection captures everything.

The trap springs when monetization is applied to decisions where multiple dimensions are active and the dimensions are not all reducible to d_1. Consider three canonical examples:

The value of a statistical life (VSL). Regulatory agencies in the United States use a VSL of approximately $11.6 million (EPA, 2023) to evaluate safety regulations. A regulation that prevents one death and costs less than $11.6 million passes the cost-benefit test; one that costs more fails. The number is derived from revealed-preference studies: how much extra pay do workers demand for jobs with higher fatality risk?

The geometric objection is not that the number is wrong but that the procedure is incoherent. Revealed-preference VSL estimates reflect d_1 trade-offs made by agents in specific labor-market contexts, with specific informational constraints (d_9), specific outside options (d_4), and specific social-norm structures (d_3, d_6). The resulting number is not a property of the value of life. It is a property of the specific manifold configuration in which the revealed preference was elicited. Transporting the number to a different policy context — say, environmental regulation affecting children who never made a labor-market choice — is parallel transporting a vector on a curved manifold. The vector arrives rotated. The number arrives wrong.

The social cost of carbon (SCC). The Interagency Working Group on the Social Cost of Greenhouse Gases sets the SCC at approximately $51 per ton of CO₂ (2020 dollars). This number enters cost-benefit calculations for every federal regulation that affects emissions. It is the monetary estimate of the damage caused by emitting one additional ton of CO₂.

The SCC compresses intergenerational harm (d_6 across decades and centuries), ecological irreversibility (d_2: rights of future populations to a livable planet), distributional injustice (d_3: poor nations bear disproportionate climate costs), and civilizational risk (d_9: epistemic uncertainty about tail outcomes) into a single dollar figure. The choice of discount rate alone — a parameter that encodes the temporal curvature of the decision manifold (Chapter 15) — swings the SCC from $14 to $200+ per ton. The “answer” depends entirely on the metric, and the metric is chosen before the calculation begins.

The compensation principle. The Kaldor-Hicks criterion declares a policy change “efficient” if the winners could, in principle, compensate the losers and still be better off — even if they do not actually compensate them. The geometric objection: the claim that the winners “could” compensate the losers is a statement about d_1 only. A policy that displaces a community (severe d_6 loss), destroys trust networks (d_5), and eliminates cultural identity (d_7) cannot be compensated by any transfer on d_1, because the losses are on dimensions that are not fungible with money. The compensation principle assumes a non-degenerate metric between all dimensions. The actual metric is degenerate: some losses cannot be offset by gains on other dimensions, at any exchange rate.

The monetization trap is, mathematically, the assumption that the metric tensor g_{\mu\nu} of the economic decision manifold is everywhere non-degenerate with finite off-diagonal components. The geometric framework makes this assumption explicit, falsifiable, and — in many important cases — false.

What Geometry Provides

The word “geometry” comes from the Greek γεωμετρία — literally, “earth measurement.” But modern geometry has traveled far from surveying. It is the mathematics of structure: how spaces are shaped, how quantities transform, how objects relate across changes of perspective.

What does this geometry provide for economics?

Directions, not just magnitudes. An economic change is not merely large or small. It points somewhere — toward greater fairness, toward less autonomy, toward more trust. A scalar captures the magnitude; a vector captures the magnitude and the direction. When we say “this policy increases welfare,” the geometric question is: in which direction? At whose expense? Along which dimensions?

A metric for comparison. To say that two values are “incommensurable” is not to say that comparison is impossible. It is to make a precise structural claim about the metric tensor: the inner product between the two value-directions is undefined, or the metric is degenerate along the relevant subspace. To say that values can be traded off at some rate is to specify a non-degenerate metric with particular off-diagonal components. Different economic theories correspond to different metrics — and the choice of metric, typically hidden in scalar frameworks, becomes explicit and debatable.

Boundaries and phase transitions. Economic life is not uniformly smooth. There are thresholds where small changes produce large jumps: the difference between being employed and unemployed, between solvent and bankrupt, between legal and illegal. Scalar functions, if continuous, cannot represent such discontinuities. Geometric economics uses stratified spaces — spaces composed of smooth regions joined along boundaries where the rules change — to represent the patchwork structure of economic reality.

Conservation laws. Emmy Noether’s theorem establishes that every continuous symmetry of a physical system corresponds to a conserved quantity. Applied to economics: the requirement that economic evaluation be invariant under re-description (the same transaction, described in different currencies, must receive the same evaluation) implies a conservation law for economic value. Value cannot be created or destroyed by relabeling. It can be generated (by production) or consumed (by use), but it must be accounted for consistently across all representations.

Computability. Geometric objects can be represented in computers, geometric operations can be implemented in algorithms, and geometric equations can be solved numerically. The companion library eris-econ demonstrates this: it implements the full decision manifold, A* pathfinding, and Bond Geodesic Equilibrium computation, and validates against published experimental data.

A unified taxonomy of error. Perhaps the most practically useful contribution of geometry is a classification of what goes wrong. In scalar economics, any deviation from rational behavior is filed under “bias” — a miscellaneous drawer that now holds over 180 entries. The geometric framework replaces this with a principled taxonomy: heuristic corruption (the price signal is inadmissible), objective hijacking (the optimizer serves the wrong master), local minima (the search is trapped in a suboptimal basin), and gauge breaking (the evaluation depends on the description, not the content). These four categories — developed in Geometric Reasoning (Bond, 2026c) and applied to economics in Chapter 10 — are structurally distinct, have different causes, different signatures, and different remedies. The geometric framework does not merely say “something went wrong”; it says which kind of thing went wrong, and why.

What This Book Is Not

Intellectual honesty requires stating what we do not claim.

This is not a claim about the metaphysics of value. We do not assert that economic dimensions exist in some Platonic realm, or that the decision manifold is a fundamental constituent of reality. Our claim is structural: that geometric representation captures more of the structure of economic phenomena than scalar representation, loses less information, and enables analysis that scalar frameworks cannot support.

This is not a new economic theory. We do not propose “geometric consequentialism” or “manifold monetarism” as competitors to existing schools. Rather, the framework provides a common mathematical language in which existing theories can be stated with unprecedented precision. Neoclassical economics is a specific metric (diagonal, d_1-only). Behavioral economics is the study of heuristic corruption on the decision manifold. Institutional economics is the study of boundary constraints. The framework does not adjudicate between these theories; it makes their commitments explicit.

This is not a rejection of mathematical economics. On the contrary, it is an argument for more mathematics — specifically, the mathematics of geometry, which provides tools that the current apparatus of optimization theory and general equilibrium lacks.

The Scalar in Practice: Three Case Studies

The scalar failure is not abstract. It shapes real decisions with real consequences every day. Three brief case studies illustrate the pattern.

The Suspicious Coincidence

In 2025, while developing a framework for verifying that AI systems treat morally equivalent inputs equivalently, I noticed something strange. The mathematics I was using to describe moral reasoning — manifolds, metrics, gauge invariance, conservation laws — was identical to the mathematics of field theory in physics. Not merely analogous. Identical.

The conservation law for harm had the same form as the conservation law for charge. The stratification of moral space had the same structure as the stratification of phase spaces. The discrete transitions between Hohfeldian jural states mapped precisely onto the D_4 dihedral group — the same symmetry group that describes the symmetries of a square.

When I applied the same framework to economic decisions, the correspondences continued. The utility function was a specific contraction of a tensor. Nash equilibrium was the scalar projection of a richer geometric object. Prospect theory’s “anomalies” — loss aversion, framing effects, the endowment effect — were natural geometric properties of a curved decision manifold with a Finsler metric.

I do not claim that this convergence reveals a deep metaphysical unity between physics and economics. Perhaps it does; perhaps it does not. What I claim is more modest and more useful: that the mathematical structures developed by physicists to describe nature are also the right structures for describing economic decision-making. The same patterns of transformation, conservation, and symmetry that organize the physical world also organize the economic world — or at least, modeling them as doing so yields a framework of extraordinary explanatory and predictive power.

The Dimensional Hypothesis

The claim of this book can be stated precisely: economic decisions have intrinsic multi-dimensional structure, and the apparent anomalies of economic behavior arise from the interaction between this structure and the scalar projections through which we observe it.

This is a testable hypothesis, not an axiom. It generates specific predictions:

Prediction: If behavioral “anomalies” are artifacts of dimensional collapse, then the anomalies should disappear when the full manifold is measured. Loss aversion (\lambda \approx 2.25) should decompose into dimensional activations: decisions that activate only d_1 should show \lambda \approx 1 (no loss aversion), while decisions that activate multiple dimensions should show \lambda > 1 (apparent loss aversion reflecting the higher manifold cost of multi-dimensional losses). Chapter 4 derives this prediction and shows that \lambda ranges from 1.0 to 3.53 depending on the number of active dimensions — exactly as the hypothesis requires.

Prediction: If the decision manifold has nine approximately independent dimensions, then factor analysis of behavioral economic data should recover approximately nine factors. Preliminary evidence from the eris-econ validation (Appendix C) supports this, but large-scale replication is needed.

Prediction: If market failures are geometric pathologies, then the four failure types (heuristic corruption, objective hijacking, local minima, gauge breaking) should be classifiable from observable market data, and the classification should predict the appropriate remedy. A market failure caused by heuristic corruption requires a different remedy than one caused by gauge breaking. Chapter 10 develops this classification.

The dimensional hypothesis is not unique to this framework. Amartya Sen’s capability approach, Kate Raworth’s Doughnut Economics, the Stiglitz-Sen-Fitoussi Commission, Bhutan’s GNH, and the OECD’s Better Life Index all assert multi-dimensionality. What the geometric framework adds is the mathematical apparatus: not just “multiple dimensions” but a metric that encodes how dimensions interact, boundaries that partition the space, conservation laws that constrain what is possible, and gauge symmetries that identify what is genuine versus what is an artifact of description.

The Arc of the Book

Part I: The Problem motivates geometric economics. This chapter has argued that scalar reduction is the structural source of economics’ most persistent failures. Chapter 2 traces proto-geometric insights through the history of economic thought — from Adam Smith’s gradient flows to Amartya Sen’s irreducibly plural capabilities — showing that the geometry was always latent in the tradition.

Part II: The Framework builds the apparatus. Chapter 3 defines the economic decision manifold. Chapter 4 equips it with a metric and derives prospect-theoretic phenomena as geometric properties. Chapter 5 develops the price signal as a heuristic field. Chapter 6 constructs the full economic tensor and proves the Scalar Irrecoverability Theorem.

Part III: Dynamics and Symmetry adds motion and conservation. Chapter 7 defines the Bond Geodesic Equilibrium and proves it subsumes Nash. Chapter 8 develops gauge invariance for economics — the principle that economic evaluations must not depend on the unit system. Chapter 9 derives conservation laws.

Part IV: Failure Modes catalogs what goes wrong. Chapter 10 maps the four geometric pathologies (heuristic corruption, objective hijacking, local minima, gauge breaking) onto market failures. Chapter 11 interprets financial crises as curvature singularities. Chapter 12 reframes inequality as metric distortion.

Part V: Applications demonstrates the framework’s reach. Chapter 13 reinterprets trade theory. Chapter 14 reframes regulation. Chapter 15 connects climate economics to manifold curvature.

Part VI: Horizons (Chapter 16) surveys open questions and the prospects for geometric economics as a research program.

A Note on Ambition

This book makes an ambitious claim: that the right mathematical language for economic decision-making is the language of modern geometry — manifolds, tensors, metrics, connections, gauge theory. This will strike some economists as overreach. Economics is a domain of practical trade-offs, institutional complexity, and irreducible uncertainty. How could it be captured by the same mathematics that describes general relativity?

The answer is that we are not claiming to capture economics. We are claiming to provide a structural vocabulary that makes economic reasoning more precise, more transparent, and more computable — without pretending to replace judgment with calculation. The vocabulary captures structure; institutional wisdom, democratic deliberation, and market dynamics fill in content. A map of a city captures the geometry of streets and landmarks; it does not capture the experience of walking through them. But the map is useful — sometimes indispensable — precisely because it captures structure that experience alone does not make explicit.

The economic map has been one-dimensional for too long. This book draws the other eight dimensions.

The Reader’s Guide

This chapter has made the case that scalar economics is structurally incomplete — that the single-number representation of economic value destroys information that is mathematically irrecoverable and practically indispensable. The remainder of the book builds the alternative.

The reader who is primarily interested in the theoretical framework should proceed to Chapter 2 (historical precursors) and then through Part II (Chapters 3–6), which builds the apparatus from the ground up. The reader who wants to see the framework in action should skip ahead to Chapter 7 (the Bond Geodesic Equilibrium) and the application chapters (13–15). The reader who is skeptical and wants to know whether any of this is empirically testable should turn to Chapter 16 (open questions and falsifiable predictions) and Appendix C (the eris-econ validation data).

Each chapter opens with a Maria’s Coffee Shop vignette that grounds the mathematics in a concrete business decision. The mathematics ranges from the accessible (Chapters 1–2, 10) to the moderately technical (Chapters 3–6, 13–15) to the formally demanding (Chapters 7–9, 11). Readers without graduate-level mathematical background may wish to read the Running Example blocks, the prose sections, and the Worked Examples while treating the formal definitions and proofs as optional. The framework’s insights are expressible in plain language; the formal apparatus exists to make them precise, not to restrict access.

A note on notation: throughout the book, d_k denotes the k-th dimension of the decision manifold (with d_1 always being the monetary/consequentialist dimension), \Sigma denotes the covariance matrix that defines the Mahalanobis metric, \beta_k denotes the boundary penalty for constraint k, and \text{BF} denotes behavioral friction — the total path cost on the full manifold. The notation is consistent across all chapters and is summarized in Appendix A.

Worked Example: Maria’s Rent Increase

Let us formalize the opening scenario. Maria’s landlord has raised the rent by 40%, from $5,000/month to $7,000/month. On the scalar projection (d_1 only), this is a simple calculation: $7,000 < revenue, so Maria can pay; $7,000 > revenue, so she closes.

On the full manifold, the decision is richer:

The Mahalanobis distance between the status quo and the post-increase state is not 40% (the scalar) but a much larger displacement on the full manifold, because multiple dimensions shift simultaneously and the covariance matrix \Sigma amplifies displacements along dimensions that co-vary (fairness and trust are correlated; when one drops, the other drops further).

The landlord’s analysis, operating on d_1 alone, sees a rent increase of $2,000/month. The geometric analysis, operating on the full manifold, sees a displacement that crosses moral boundaries (d_3, d_7), degrades trust (d_5), destroys social capital (d_6), and creates existential uncertainty (d_9). The boundary penalties \beta_k for the crossed thresholds dominate the Mahalanobis distance. The “rational” d_1 decision — raise rent to market — is revealed as a suboptimal path on the full manifold.

This is not a rejection of the landlord’s right to charge market rent. It is a demonstration that the market rate itself is a scalar projection of a richer economic reality, and that decisions made on the projection can be strictly dominated by decisions made on the full manifold — dominated in the Pareto sense, where both landlord and community are better off under a path that respects the full geometry.

Technical Appendix

The Scalar Irrecoverability Theorem (Preview). [Conditional Theorem.] Let X \subseteq \mathbb{R}^n be a connected open set with n \geq 2. No continuous function \phi: X \to \mathbb{R} is injective. Therefore, any scalar summary of a multi-dimensional economic state necessarily maps distinct states to the same number, and no post-hoc procedure can recover the lost distinction.

Proof sketch. By Brouwer’s invariance of domain, \mathbb{R}^n and \mathbb{R}^1 are not homeomorphic for n \geq 2. If \phi were a continuous injection from the compact set \overline{X} into \mathbb{R}, it would be a homeomorphism onto its image — embedding an n-dimensional space into \mathbb{R}^1. This contradicts invariance of domain. The full proof is given in Chapter 6.

Dimensional Count (Justification). [Modeling Axiom.] The claim that economic decisions involve nine dimensions is a modeling axiom, not a mathematical necessity. The specific dimensionality is inherited from the moral manifold developed in Geometric Ethics (Bond, 2026b, Chapter 5), where it is derived from a 3 \times 3 decomposition of moral space (three normative modes × three scopes). The adaptation to economics replaces moral-specific labels with economic-specific ones while preserving the mathematical structure. The dimensionality is empirically testable: factor analysis of behavioral economic data should recover approximately nine independent factors. The eris-econ validation (Appendix C) provides preliminary evidence.

The Contraction Operation. [Established Mathematics.] A contraction is a linear map from a tensor to a lower-rank tensor, obtained by summing over a pair of indices. GDP is a contraction of the economic tensor: it sums over all agents i, all goods j, and all non-monetary dimensions k > 1, retaining only the monetary value of final transactions. In index notation:

\text{GDP} = \sum_{i} \sum_{j} T^{1}_{ij}

where T^k_{ij} is the economic tensor with dimension index k, agent index i, and good index j. The contraction discards the k > 1 components (non-monetary dimensions) and the off-diagonal structure (how dimensions interact). The Scalar Irrecoverability Theorem (Chapter 6) proves that no linear operation on GDP can recover the discarded components — the contraction is mathematically irreversible.

The Mahalanobis Distance (Preview). [Established Mathematics.] The Mahalanobis distance between two points \mathbf{x} and \mathbf{y} in a space with covariance matrix \Sigma is:

d_M(\mathbf{x}, \mathbf{y}) = \sqrt{(\mathbf{x} - \mathbf{y})^T \Sigma^{-1} (\mathbf{x} - \mathbf{y})}

This is the natural distance measure when the dimensions have different scales and are correlated. When \Sigma is diagonal, the Mahalanobis distance reduces to a weighted Euclidean distance with weights 1/\sigma_k^2. When \Sigma has off-diagonal terms, the distance depends on the correlations between dimensions — a displacement that is aligned with the correlation structure costs less than one that is orthogonal to it. Chapter 4 develops the full economic metric based on this distance measure.

Notes on Sources

The critique of GDP as a welfare measure is developed in Stiglitz, Sen, and Fitoussi (2009), Mismeasuring Our Lives. Raworth (2017), Doughnut Economics, provides the most accessible recent critique from within the economics profession. The Scalar Irrecoverability Theorem is a mathematical formalization of their qualitative arguments.

The behavioral economics canon includes Kahneman and Tversky (1979), “Prospect Theory: An Analysis of Decision Under Risk,” Econometrica; Thaler (1980), “Toward a Positive Theory of Consumer Choice,” Journal of Economic Behavior and Organization; and Ariely (2008), Predictably Irrational. We treat these as empirical data, not as theoretical commitments.

The geometric framework inherits from Bond (2026a), Geometric Methods in Computational Modeling; Bond (2026b), Geometric Ethics; and Bond (2026c), Geometric Reasoning: From Search to Manifolds. The specific application to economics was first presented as Chapter 20 of Geometric Ethics and is developed here in full.

The Easterlin paradox is documented in Easterlin (1974), “Does Economic Growth Improve the Human Lot?”, in Nations and Households in Economic Growth (Academic Press), and updated in Easterlin et al. (2010), “The Happiness-Income Paradox Revisited,” PNAS. The geometric interpretation — that the paradox reflects the decoupling of the correlation structure between monetary and evaluative dimensions above a threshold — is original to this framework.

The monetization trap is discussed, under different names, in Anderson (1993), Value in Ethics and Economics (Harvard University Press); Sandel (2012), What Money Can’t Buy: The Moral Limits of Markets (FSG); and Satz (2010), Why Some Things Should Not Be for Sale (Oxford University Press). The geometric formalization — that the trap is the assumption of a non-degenerate metric where the true metric is degenerate — is original.

The value of a statistical life (VSL) methodology is reviewed in Viscusi (2018), Pricing Lives: Guideposts for a Safer Society (Princeton University Press). The social cost of carbon is reviewed in the Interagency Working Group on Social Cost of Greenhouse Gases (2021). Bhutan’s Gross National Happiness is documented in the Centre for Bhutan Studies (2012), GNH Survey Report. The Amazon logistics case draws on the reporting compiled in Kantor and Streitfeld (2015, New York Times) and subsequent investigations. The opioid crisis statistics are from the CDC’s National Center for Health Statistics.