we-solve (beta)

Econ-479 · economics · Economics

Serial acquisitions

LLM status Open unverified

Statement

Original subject group: Industrial organization and competition

Audit classification: Background; proposed extension. The source verifies a policy concern about serial acquisitions; it does not publish this formal cumulative-price question. Isolated-deal effects must be distinguished from telescoping sequential increments.

Definitions and precise research target

Fix local United States outpatient-provider markets and a ten-year observation window. For each acquirer $m$, let $A_m=(a_{m1},\ldots,a_{mK_m})$ be its observed sequence of acquisitions, where $K_m$ is the number of deals and every deal is individually below the contemporaneous notification threshold. The threshold is a descriptive selection rule, not an assumption of competitive harmlessness. Define $P_m(A_m,h)$ as the equilibrium price index at horizon $h=5$ years under that acquisition sequence, permitting rival entry, exit and investment. Define $P_m(0,h)$ under the counterfactual sequence containing no acquisitions by that acquirer, with other institutional changes specified identically. The target cumulative effect is $\tau(h)=E[P_m(A_m,h)-P_m(0,h)]$, averaged over this below-threshold acquirer population. Compare it with the sum of effects attributed to the individual deals to measure nonadditivity. Construct ownership histories covering unreported transactions and address endogenous choice of targets and acquisition timing. A valid comparison requires stated counterfactual-trends or instrument assumptions, with bounds when these assumptions are uncertain. Report heterogeneity by baseline concentration and overlapping service lines. This agenda quantifies cumulative effects in one service sector; it does not infer an effect from notification status alone or claim that all small acquisitions reduce competition.

Definitions, assumptions and mathematical qualifications. The input sequence $A_m$ includes each acquired target, closing date and acquired service line. Its zero counterpart holds all other policy rules fixed but permits endogenous responses. Define isolated-deal sequence $A_m^{(k)}$ as only deal $k$, at its observed date, and use a common terminal calendar date. The nonadditivity contrast is $\mathcal N_m=P_m(A_m,h)-P_m(0,h)-\sum_k\{P_m(A_m^{(k)},h)-P_m(0,h)\}$. This is different from the sum of sequential marginal effects, which telescopes to the total by construction. Eligibility uses the notification rule in force on each actual transaction date; historical thresholds must be independently sourced rather than replaced by today's threshold.

Variants and assumption changes

  1. Editorial, additive benchmark. Under $P_m(A)=P_m(0)+\sum_{k\in A}\gamma_{mk}$, $\mathcal N_m=0$ algebraically; estimate departures rather than call additivity a general open theorem.
  2. Editorial, endogenous histories. Model acquisition timing and targets as part of a dynamic structural law or use a specified quasi-experiment. Report bounds over admissible histories when observed transactions and selection mechanisms are incomplete.

References and source audit

Checked source. 24 August 2024 author draft, Section 6, printed p. 35, serial-acquisition policy recommendation. Full primary text retrieved. https://web.stanford.edu/~ayurukog/trendsincompetition.pdf.
  1. Carl Shapiro; Ali Yurukoglu. 2024. Trends in Competition in the United States: What Does the Evidence Show?. 24 August 2024 author draft, Section 6, printed p. 35, serial-acquisition policy recommendation. https://web.stanford.edu/~ayurukog/trendsincompetition.pdf. DOI: https://doi.org/10.3386/w32762.

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-479-A1 · Exploration · LLM-generated · gpt 6 astra ultra ·

GPT 6 Astra Ultra attempt on Econ-479

Serial acquisitions: exact ownership interactions and an entry-induced sign reversal

Problem ID: L41-475. Status: Unresolved; rigorous restricted results.

Target and benchmark

The target compares a complete below-notification-threshold acquisition history with no acquisitions, and compares that effect with the sum of isolated deals at the same terminal date. Shapiro and Yurukoglu [sy], Section 6, printed p.35, motivate attention to serial acquisitions. Their policy discussion does not establish the cumulative-price formula studied here.

We derive exact cumulative effects in a homogeneous Cournot ownership model, then show that allowing entry can reverse the sign of nonadditivity. We also identify which experimental cells are needed to recover general ownership interactions. This restricted benchmark has no investment dynamics, quality heterogeneity, or endogenous target choice. It therefore does not identify the five-year effect in actual outpatient markets. Its advantage is that the counterfactual price equilibria and the entry response are fully specified.

Ownership consolidation without entry

Inverse demand is $p(Q)=a-bQ$, with $a>c\geq0$ and $b>0$. Initially $N\geq2$ independently owned firms produce a homogeneous service at constant marginal resource cost $c$, without capacity constraints. Firms choose nonnegative quantities simultaneously. A firm acquiring a rival centrally chooses the combined organization's total quantity; it does not commit to preserving each former establishment's independent output decision. No efficiency gain occurs. A serial acquirer buys $K$ distinct rivals, where $0\leq K\leq N-1$. All counterfactuals are evaluated after quantity adjustment at one common terminal date.
Lemma (Unique quantity equilibrium): With $n\geq1$ independent owners, the unique equilibrium has \[ q_j(n)=\frac{a-c}{b(n+1)},\quad Q(n)=\frac{n(a-c)}{b(n+1)},\quad p(n)=c+\frac{a-c}{n+1}. \tag{1} \] Each owner's operating profit is $(a-c)^2/[b(n+1)^2]$.
Proof: Given rivals' total $Q_{-j}$, profit is $(a-c-bQ_{-j})q_j-bq_j^2$, a strictly concave quadratic. At an equilibrium an active firm's first-order condition is $a-c-bQ-bq_j=0$, so all active firms have the same output and $q_j=(a-c)/[b(m+1)]$ if $m$ firms are active. If any firm were inactive, its marginal profit at zero would be $a-c-bQ=(a-c)/(m+1)>0$, a contradiction. Thus $m=n$, establishing (1) and uniqueness. Substitution gives operating profit. At these outputs price exceeds $c$, so the solution does not rely on negative prices. ∎

Let $d=a-c$. The cumulative price effect is $\tau_K=p(N-K)-p(N)$; every isolated acquisition leaves $N-1$ owners and has effect $\tau_1$. Define $\mathcal N_K=\tau_K-K\tau_1$.

Proposition (Strict superadditivity of ownership removal): For $0\leq K\leq N-1$, \[ \tau_K=\frac{dK}{(N+1)(N-K+1)},\qquad \mathcal N_K=\frac{dK(K-1)}{N(N+1)(N-K+1)}. \tag{2} \] Thus isolated deals understate the total for $K\geq2$.
Proof: Equation (1) gives $\tau_K=d\{(N-K+1)^{-1}-(N+1)^{-1}\}$ and $\tau_1=d/[N(N+1)]$. Subtracting $K\tau_1$ and taking a common denominator gives (2). All denominators are positive. For $K=0,1$ the numerator of $\mathcal N_K$ is zero, and for $K\geq2$ it is positive. ∎

This is an ownership result, not a theorem about physical capacity destruction. The merged owner can produce at all its establishments at cost $c$. It internalizes their competition. The history is an externally specified intervention; the proposition does not claim that every purchase is privately profitable at an equilibrium acquisition price. Historical notification thresholds select the empirical cohort but do not enter (1) or establish harmlessness.

Sequential marginal effects are $p(N-k)-p(N-k+1)$ and sum to $\tau_K$ by cancellation. Consequently comparing the total to that sequential sum cannot test nonadditivity. Equation (2) uses the genuinely different counterfactual in which each deal occurs alone.

Entry can reverse the conclusion

Now let the $n_0=N-K$ incumbent owners have sunk their fixed entry costs. There are $M$ potential entrants, each with fixed resource cost $F>0$ and the same marginal cost $c$. Potential entrants simultaneously choose whether to enter; all active firms then play the unique quantity equilibrium (1). Incumbents cannot exit at this terminal stage. Entry tie-breaking favors entry at zero profit. Define \[ \pi(n)=\frac{d^2}{b(n+1)^2},\qquad n_* =\max\left(\{n\geq1:\pi(n)\geq F\}\cup\{0\}\right). \] The integer is finite. Assume $M\geq\max(0,n_*-n_0)$ and enough additional candidate firms exist to evaluate entry deviations when needed.
Lemma (Selected entry equilibrium): An equilibrium can be selected with total active firms \[ n(K)=\max\{N-K,n_*\}. \tag{3} \] When several entrants are interchangeable, selecting the lowest indexed entrants fixes their identities. If $\pi(n_*)=F$, the specified zero-profit entry convention selects (3).
Proof: If $n_0\geq n_*$, no entry occurs. An entrant would then earn $\pi(n_0+1)-F<0$, by maximality of $n_*$. If $n_0<n_*$, select exactly $n_*-n_0$ entrants. Each earns $\pi(n_*)-F\geq0$ and does not gain by exit; any additional entrant would earn $\pi(n_*+1)-F<0$. This verifies every entry-stage unilateral deviation. The second-stage quantity equilibrium is unique, so these entry choices and continuation quantities form a subgame-perfect outcome. At zero-profit ties another count can also be an equilibrium; the declared selection chooses the maximal count permitted by (3). ∎

The convention "enough additional candidate firms" only matters for proving a no-entry inequality, not for attaining the stated count; a finite $M\geq n_*$ suffices uniformly over all histories. Assume $N\geq n_*$ so that the baseline has no new entry. Terminal price is \[ P(K)=c+\frac{d}{\max\{N-K,n_*\}+1}. \tag{4} \] It rises with acquisitions until entry fully replaces further removed owners, and is then flat.

Theorem (Opposite signs under different entry costs): For $N\geq3$ and $2\leq K\leq N-1$, both positive and negative nonadditivity are possible in this optimizing model. If $n_*\leq N-K$, the effect is the positive expression (2). If $n_*=N-1$, then \[ P(K)-P(0)=\frac{d}{N(N+1)},\qquad \mathcal N_K=(1-K)\frac{d}{N(N+1)}<0. \tag{5} \] The latter case is attained by every $F$ satisfying \[ \frac{d^2}{b(N+1)^2}<F\leq\frac{d^2}{bN^2}. \]
Proof: The first condition prevents entry in all histories with at most $K$ acquisitions, so (2) applies. Under the second condition every nonempty acquisition history induces exactly $N-1$ terminal firms: one removed owner is not replaced, but every further removal induces an entrant. Therefore all isolated-deal and bundled price increments equal $d/[N(N+1)]$, proving (5). The displayed interval for $F$ is nonempty and says precisely $\pi(N)<F\leq\pi(N-1)$, equivalent to $n_*=N-1$. Very large $F>\pi(1)$ instead gives $n_*=0$ and the no-entry case. ∎

This reversal uses the same demand and marginal cost throughout. It shows why concentration measured immediately after each deal cannot replace the selected equilibrium after entry. It also provides a falsifiable restricted prediction: once total ownership hits the entry threshold, incremental acquisitions generate one-for-one entrants and no further price increase. Investment or differentiated services can invalidate this threshold mechanism; that is an empirical limitation rather than a missing step in (5).

The experimental meaning of interactions

For a finite catalogue $\mathcal K=\{1,\ldots,K\}$ of dated deals, let $F(S)$ be expected terminal price when exactly the subset $S$ occurs, with their original dates retained. Define, for every $T\subseteq\mathcal K$, \[ \beta_T=\sum_{U\subseteq T}(-1)^{|T|-|U|}F(U). \]
Proposition (Interaction decomposition and identification): For every $S$, $F(S)=\sum_{T\subseteq S}\beta_T$, and \[ F(\mathcal K)-F(\varnothing) -\sum_{k=1}^K\{F(\{k\})-F(\varnothing)\} =\sum_{|T|\geq2}\beta_T. \tag{6} \] Random assignment across the empty, full, and all singleton histories identifies (6), without identifying each higher-order interaction separately. Assignment only between the full and empty histories generally does not identify (6).
Proof: In $\sum_{T\subseteq S}\beta_T$, the coefficient of $F(U)$ is $\sum_{U\subseteq T\subseteq S}(-1)^{|T|-|U|}=(1-1)^{|S\setminus U|}$. It is one for $U=S$ and zero otherwise. This proves inversion and then (6) by subtracting constant and singleton terms. Under random assignment of whole noninterfering markets and consistent implementation, each assigned history's conditional mean equals its $F(S)$, which proves the identification claim. To prove the last statement, keep $F(\varnothing)$ and $F(\mathcal K)$ fixed and change any singleton value when $K\geq2$. This changes (6) while leaving both observed arms unchanged. Such finite arrays are realizable as deterministic potential-price models, so nonidentification is substantive in the unrestricted history-response class. ∎

Finally, suppose administrative reconstruction only establishes $K\in\{k,k+1,\ldots,k+u\}$, and $d,N,n_*$ are known with $k+u\leq N-1$. In the count-only model (4), the sharp cumulative-effect set is the finite collection $\{P(j)-P(0):j=k,\ldots,k+u\}$. Its smallest containing interval has endpoints at $k$ and $k+u$ because (4) is nondecreasing. Calling every interior point sharp would be incorrect unless the target averages over a population with unrestricted mixtures of counts. Under that mixture interpretation the convex interval is sharp, by mixing the two endpoint counts.

The remaining work is to identify demand, cost, entry response and the acquisition-selection law in the specified provider cohort, not merely to fit an event-study coefficient for each observed transaction. No historical threshold values or empirical price effects are asserted here.

Completion Estimate

45%

This is a post hoc, heuristic estimate for the full catalogue target. The Cournot ownership model proves exact cumulative superadditivity and an optimizing entry mechanism that reverses nonadditivity, with the necessary experimental history cells identified. Actual ownership reconstruction, endogenous targets and timing, defensible counterfactual trend bounds, investment and quality responses, and five year outpatient price effects remain unestablished.

References

Go to contribution →