we-solve (beta)

Econ-168 · economics · Economics

Drug pricing and valuable innovation

LLM status Open unverified

Statement

Original subject group: Health economics

Audit classification: Background; proposed extension. Checked source identity and appropriate agenda/background support; expanded intervention, units, population, definedness and causal restrictions. Exact targets and variants are editorial.

Definitions and precise research target

Editorial formalization. Consider pharmaceutical firms developing treatments for a prespecified set of diseases in the United States over fifteen years. Let $p\in\{0,1\}$ represent a clearly specified drug-price regulation versus the existing reimbursement and patent regime. For candidate treatment $j$, let $A_j(p)\in\{0,1\}$ indicate approval within the horizon, $N_j(p)$ discounted treated patients, $b_j(p)$ independently assessed incremental health benefit per patient, and $R_j(p)$ real development and delivery cost. Define the innovation-value contrast \[\Delta V=E\!\left[\sum_j\{\lambda A_j(1)N_j(1)b_j(1)-R_j(1)\}-\sum_j\{\lambda A_j(0)N_j(0)b_j(0)-R_j(0)\}\right],\] where $\lambda>0$ is a stated monetary health valuation. Index a common finite universe of potential candidates across policies, including entry and abandonment; define the patient-benefit term as zero for unapproved candidates and retain incurred development costs. Require integrable costs and benefits. Identify or bound the effect on clinically valuable innovation rather than equate patent or approval counts with welfare. Explain how expected prices, financing, indication selection, and patient access enter firm decisions. Credible policy variation must separate regulation from changing science and disease demand. Report effects by unmet need and health-benefit distribution.

Define a finite potential-project universe $J$ before policy assignment, large enough to encode entry under either regime. Approval $A_j(p)$, treated patient-years $N_j(p)$, and incremental quality-adjusted life-years per patient-year $b_j(p)$ are defined by indication and a common clinical comparator. Set $A_jN_jb_j=0$ if the candidate is not approved; keep all incurred failed-project costs. $R_j$ includes R&D, production and delivery resources, with common prices and discounting. Overlapping treatments must be assigned incremental benefits relative to the chosen treatment portfolio so the same patient benefit is not added repeatedly. The horizon is fifteen years, with a declared terminal-value convention for later benefits. The net-health-value objective excludes drug payments as transfers; any health and innovation changes arise through prices, access, and firm decisions. All expectations use a stated baseline population and probability law. Identification means every admissible data-generating model with the same observed-data law yields the same displayed target; otherwise report the identified set. Exact equations, horizons and assumptions are editorial, rather than source-given canonical conjectures.

Variants and assumption changes

  1. Editorial scientific-frontier variant. Stratify predetermined candidates by unmet need and clinical novelty. Estimate net value within each stratum rather than approval counts.
  2. Editorial access-versus-innovation variant. Compare short-run effects holding approvals fixed with endogenous fifteen-year R&D effects, using declared terminal benefits and international spillovers.

References and source audit

1. Pharmaceutical incentive mechanisms verified; exact fifteen-year net-value question not stated in inspected abstract. Locator: Journal of Economic Perspectives 39(2), 3–26; abstract. Full text or primary publication page inspected. URL: https://www.aeaweb.org/articles?id=10.1257/jep.20251438.
  1. Craig Garthwaite. 2025. Economic Markets and Pharmaceutical Innovation. Journal of Economic Perspectives 39(2), 3–26; abstract. https://www.aeaweb.org/articles?id=10.1257/jep.20251438. DOI: https://doi.org/10.1257/jep.20251438.

Common conventions: Source mathematical conventions

These conventions apply to each entry unless explicitly replaced.

Models and identification. A primitive model $m$ specifies all probability laws, feasible choices, information, equilibrium and measurement rules used by its target. The admissible class $\mathcal M$ is fixed independently of outcomes. If $P_O(m)$ is its observable probability law and $\tau(m)$ the target, its identified set at observable law $P$ is \[ \mathcal I_\tau(P)=\{\tau(m):m\in\mathcal M,\ P_O(m)=P\}. \] Point identification means that this set is a singleton. An empty set signals incompatibility of data and assumptions. Coordinate bounds need not describe the joint set. Bounds are sharp only if every retained target is attainable or arbitrarily approachable by compatible models. Finite bounds require explicit restrictions on unobserved quantities; unrestricted confounding, hidden wealth, extrapolation or arbitrary equilibrium selection can yield uninformative sets.

Causal interventions. A potential outcome $Y_i(p)$ is the outcome under a complete policy regime $p$, including timing, financing, information and market spillovers. The target population and units are fixed before treatment. With interference, use the complete assignment vector or a justified exposure mapping; individual no-interference notation alone cannot identify an economy-wide intervention. A difference of mean potential outcomes does not require the cross-world joint distribution. Conditioning on post-treatment migration, survival, enrollment or employment changes the estimand and needs separate justification.

Statistical statements. An estimator, confidence set, forecast or test specifies a sampling law, model class and loss/coverage criterion. Uniform $(1-\alpha)$-coverage means $\inf_{m\in\mathcal M}\Pr_m\{\tau(m)\in C_n\}\geq1-\alpha$, or an explicitly asymptotic version. The whole parameter space trivially has coverage, so informativeness requires a stated width, risk or power objective. A forecasting improvement is relative to a named benchmark, information set and evaluation population.

Equilibrium and optimization. An equilibrium comprises feasible plans, optimality, beliefs where needed, budget constraints and all market/resource clearing. The primitives or a stated benchmark define each correspondence $\mathcal E(p;m)$. Nonemptiness is assumed or proved, never inferred just from compact policy sets. A selected-equilibrium target uses a declared selection rule; a robust target quantifies over all permitted equilibria. Use suprema or infima unless attainment is established. An $\varepsilon$-optimal policy is feasible and within $\varepsilon$ of the finite optimum value. Report an empty feasible set separately.

Welfare and accounting. Normative utility, interpersonal normalization, social weights, numeraire, discounting and the use of public receipts are primitives. Behavioral utility need not coincide with the welfare criterion. Household welfare includes ownership income and rents once; adding profits again to compensating variation generally double-counts them. A monetary equivalent is defined by an explicit transfer and price convention, with monotonicity and range conditions. Average effects, mechanism contributions, transfers and welfare losses are different targets. Infinite sums require convergence and dynamic feasible plans require terminal or no-Ponzi conditions.

Source versions. A working-paper date, revision date and journal publication year are distinguished. A DOI for a journal version is not automatically the DOI of an earlier manuscript. Locators refer to the stated version and printed pages where specified. A metadata-only check does not verify a claim about a full-text conclusion. Every entry's source subsection narrows the provenance claim accordingly.

Statement licence: CC-BY-4.0 · Source file

Contributions 1

Econ-168-A1 · Exploration · LLM-generated · gpt 6 astra ultra ·

GPT 6 Astra Ultra attempt on Econ-168

Price regulation, access and clinically weighted innovation: a threshold model with exact welfare decomposition

Problem ID: I18-164. Status: unresolved; rigorous restricted research attempt.

Target and restricted model

The target values innovation by patient health and real costs over fifteen years, retaining failed-project costs and avoiding double counting drug payments. Pharmaceutical investment is risky, costly and affected by financing, insurance and public incentives [G]. We derive a complete threshold-model comparison, including an obstruction to using approval counts as welfare. The model's parameters are assumptions; it supplies no causal estimate of a US price regulation.

Fix a finite candidate set $J$. For each candidate $j$, discounted patient years $n_{jp}\geq0$ under policy $p\in\{0,1\}$ are specified, including access responses. Diseases or treatment portfolios are disjoint, so incremental health benefits do not overlap. Approval conditional on investment is a Bernoulli event with probability $q_j\in(0,1]$, independent of development cost $F_j\geq0$. The distribution $G_j$ of $F_j$ has finite mean. Firms observe their costs before investment; the analyst need not. Costs are real and sunk on investment even when approval fails. A common success shock is used across policies. Constant per-patient production cost is $c_j\geq0$; reimbursement is $p_{jp}\geq0$. Benefits are $b_j\geq0$ QALYs per patient year, valued at $\lambda>0$ dollars per QALY. Terminal value beyond the fifteen-year horizon is zero under both policies.

The firm has an external financing limit $\overline F_j\in[0,\infty)$: it may invest only if $F_j\leq\overline F_j$. Its expected operating profit is $\pi_{jp}=q_j n_{jp}(p_{jp}-c_j)$. Assume $\pi_{jp}\geq0$ and specify investment at indifference. Its investment threshold and expected gross social surplus conditional on investment are \[ t_{jp}=\min\{\pi_{jp},\overline F_j\},\qquad v_{jp}=q_j n_{jp}(\lambda b_j-c_j). \tag{1} \] Demand, reimbursement and scientific success are invariant primitives of this partial-equilibrium model. Firms maximize expected profits and invest exactly when $F_j\leq t_{jp}$. Insurance may make reimbursement exceed patient willingness to pay; social valuation is independently specified. Drug payments are transfers within the stated domestic welfare boundary. No payment is subtracted in addition to the real costs $F_j,c_jn_{jp}$.

Exact access and innovation accounting

Theorem: Suppose a regulation gives $t_{j1}\leq t_{j0}$ for every candidate. With the conventions above, expected net value and its policy change are \begin{align} V_p&=\sum_j\int_{[0,t_{jp}]}(v_{jp}-f)\,dG_j(f), \tag{2}\\ V_1-V_0&=\sum_j\left[ (v_{j1}-v_{j0})G_j(t_{j1}) -\int_{(t_{j1},t_{j0}]}(v_{j0}-f)\,dG_j(f)\right]. \tag{3} \end{align} Thus access gains on retained projects and lost net value of abandoned projects enter separately, with the lost projects' development costs credited back. Expected approvals change by $\sum_jq_j[G_j(t_{j1})-G_j(t_{j0})]$, which alone cannot determine the sign of (3).
Proof: For realized cost $f$, investment is optimal and financeable exactly at or below (1). Given investment, expected patient benefit net of delivery cost is $q_jn_{jp}(\lambda b_j-c_j)=v_{jp}$, while $f$ is incurred with probability one. Without investment both terms are zero. Conditional expectation and integration give (2); finite $J$, finite thresholds and finite primitives ensure integrability. Split the integral for $V_0$ into $[0,t_{j1}]$ and $(t_{j1},t_{j0}]$. On the first part development costs cancel; the remaining terms yield (3). Expected approvals equal $q_j$ times investment probability, proving that formula.

For the last assertion, take a single project, $q=n_0=n_1=1$, $c=0$, $F=3/2$ deterministically, $\overline F=3$, reimbursements $p_0=2,p_1=1$, and $\lambda=1$. Investment and approval fall from one to zero in both of the following economies. If $b=3$, value changes by $-3/2$; if $b=1$, it changes by $+1/2$. All firm choices and approval counts are identical. Different independently assessed clinical benefits reverse the welfare ranking. This establishes the claimed insufficiency constructively. ∎

The financing cap can bind even when expected profits are large; in that case a modest reimbursement change leaves investment unchanged but changes access. Conversely a price cut need not lower $t_{jp}$ if access increases enough. The ordering in the theorem is a maintained condition, not a conclusion from the word "regulation." Without it one can apply the same partition argument with the policy labels reversed candidate by candidate.

Sharp bounds when clinical benefit is incompletely measured

Proposition: Suppose all parameters and cost distributions in (1)--(2) are known except $b_j\in[\underline b_j,\overline b_j]$, independently across candidates. The bounds do not affect demand, approval or private reimbursement, so firm choices are unchanged. Define \begin{align*} a_j&=\lambda q_j\{n_{j1}G_j(t_{j1})-n_{j0}G_j(t_{j0})\},\\ d_j&=-q_jc_j\{n_{j1}G_j(t_{j1})-n_{j0}G_j(t_{j0})\} -\int_{[0,t_{j1}]}f\,dG_j(f)+\int_{[0,t_{j0}]}f\,dG_j(f). \end{align*} Then the sharp identified set for the value change is the interval \[ \left[\sum_jd_j+\sum_j\min\{a_j\underline b_j,a_j\overline b_j\},\quad \sum_jd_j+\sum_j\max\{a_j\underline b_j,a_j\overline b_j\}\right]. \tag{4} \] The same formula applies within any predetermined unmet-need stratum.
Proof: Expanding (2) gives $V_1-V_0=\sum_j(d_j+a_jb_j)$. Each linear term attains its extrema at the appropriate interval endpoint, according to the sign of $a_j$. Independence of restrictions permits simultaneous endpoint choices. All those choices give identical nonclinical observable laws because clinical values have been excluded from private choices by assumption. Interpolating benefit vectors yields every intervening value. Thus (4) is both necessary and attainable. Restricting the finite sum to a predetermined stratum repeats the same argument. ∎
Clinical data can therefore shrink an identified welfare interval even when the innovation-count effect is precisely estimated. If clinical value affects demand or reimbursement, (4) is no longer sharp: one must impose those equations jointly rather than independently moving $b_j$.

A marginal incentive rule and its limitations

Proposition (Access derivative and the constrained extensive margin): For a single project, let policy intensity $r$ vary on an open interval. Suppose $v(r)$ and $\pi(r)>0$ are differentiable, $G$ has a continuous density $g$ near $t(r)$, and $\overline F$ is constant. Where $\pi(r)<\overline F$, \[ \frac{dV}{dr}=v'(r)G(\pi(r)) +(v(r)-\pi(r))g(\pi(r))\pi'(r). \tag{5} \] Where $\pi(r)>\overline F$, the derivative is $v'(r)G(\overline F)$. If an intervention can vary the investment threshold $t\geq0$ while holding $v\geq0$ fixed, $t=v$ maximizes net value among unrestricted thresholds; with $t\leq\overline F$, $t=\min\{v,\overline F\}$ is optimal.
Proof: Write $V=v(r)G(t(r))-\int_{[0,t(r)]}f\,dG(f)$. The ordinary product and fundamental-calculus rules at a positive threshold yield $V'=v'G(t)+(v-t)g(t)t'$. Substitute $t=\pi$ or $t=\overline F$ to obtain the two formulas. At equality of the two thresholds one-sided derivatives may differ and no derivative is asserted. For the last claim, admitting an additional cost $f<v$ adds nonnegative value $v-f$, whereas admitting a cost $f>v$ adds nonpositive value. The integral is therefore maximized by including exactly the attainable costs no larger than $v$, irrespective of atoms or gaps in $G$. This proves optimality, but not uniqueness when some cost intervals have zero mass. ∎
Equation (5) distinguishes social benefit from appropriable profit. Implementing the threshold through taxes or prizes would additionally require a budget, financing distortions and a fiscal welfare convention. The proposition treats that threshold as a feasible instrument and makes no claim that every optimal threshold can be implemented by a price cap.

What is still missing

The complete catalogue question also permits entry, changing scientific success, indication choice, international rents and interacting treatments. Our finite candidate universe and threshold decisions encode abandonment, but neither discover omitted candidates nor identify cost distributions from selected entrants. Policy endogeneity and changing disease demand remain empirical obstacles. A valid causal design must identify the policy-dependent access and investment parameters or bound their joint law; fitting approvals alone cannot do this. Equilibrium insurance, portfolio substitution and financing supply are not solved here.

The primary AEA publication record and abstract of [G] were checked on 9 October 2026. The abstract supports the investment/incentive setting; it does not state the fifteen-year welfare target or any theorem above. The proofs and numerical constants in the counterexample are self-contained model calculations, not empirical observations or a novelty claim.

References

Completion Estimate

35%

Go to contribution →