Protection for Sale Model

Protection for Sale Model is the workhorse political-economy theory of endogenous trade policy in international economics and business and economics theory, developed by economists Gene M. Grossman and Elhanan Helpman in their landmark 1994 American Economic Review article "Protection for Sale." Moving beyond earlier reduced-form political support models and tariff-formation lobbying contests, Grossman and Helpman modeled the interaction between organized special-interest industry lobbies and an incumbent government as a common agency (menu auction) game based on B. Douglas Bernheim and Michael D. Whinston (1986). In a small open specific-factors economy, owners of sector-specific capital in an exogenous subset of industries i ∈ L overcome collective-action frictions and offer continuous, policy-contingent campaign contribution schedules C_i(p) that map the vector of domestic trade-tax-distorted prices p = (p₁, ..., p_n) into political contributions, while the incumbent government selects p to maximize a weighted sum of total campaign contributions and aggregate social welfare: G = ∑_i∈L C_i(p) + a W(p). Grossman and Helpman proved that in any truthful subgame-perfect Nash equilibrium, the equilibrium ad valorem tariff (or export subsidy) in each sector i obeys a testable Modified Ramsey Pricing Formula: (t_i / (1 + t_i)) = [ (I_i − α_L) / (a + α_L) ] × ( z_i / e_i ), predicting that organized industries (I_i = 1) receive positive import protection or export subsidies while unorganized industries (I_i = 0) are taxed, and that among organized industries protection increases with the ratio of domestic output to imports (z_i = X_i / M_i) and decreases with the price elasticity of import demand (e_i).

Production, Factor Incomes, and Quasi-Linear Consumer Welfare

Consider a small open economy populated by a continuum of citizens of total mass normalized to 1. Each individual has identical, additively separable, quasi-linear preferences over a numeraire good x₀ (whose world and domestic price are normalized to 1) and n non-numeraire tradable goods i = 1, 2, ..., n:

where each sub-utility function u_i(·) is differentiable, increasing, and concave. Facing domestic price p_i for good i, every consumer demands quantity d_i(p_i) = (u′_i)⁻¹(p_i) and achieves indirect utility I + s(p), where I is disposable income, p = (p₁, ..., p_n) is the domestic price vector, and s(p) is aggregate consumer surplus:

On the production side, the numeraire good 0 is produced using labor alone under constant returns to scale with a unit input-output coefficient, pinning the economy-wide competitive wage rate to 1 (w = 1). Each non-numeraire sector i ∈ {1, ..., n} produces output using mobile labor (at wage w = 1) and an inelastic sector-specific factor (industry-specific capital or land) under constant returns to scale. Given wage w = 1, the restricted profit (rent) accruing to the owners of sector-i specific capital is an increasing, convex function Π_i(p_i), whose derivative by Hotelling's Lemma equals domestic industry supply X_i(p_i):

Trade Instruments, Tariff Revenue, and Aggregate Social Welfare

Let p_i^* denote the exogenous world price of good i (Grossman and Helpman, 1995, and the Bagwell–Staiger terms-of-trade theory extend the framework to large countries). The government controls trade taxes and subsidies that create wedges p_i − p_i^* between domestic and world prices:

• For an import-competing sector (M_i(p_i) ≡ d_i(p_i) − X_i(p_i) > 0), p_i > p_i^* is an import tariff (and p_i < p_i^* is an import subsidy).
• For an exporting sector (M_i(p_i) < 0), p_i > p_i^* is an export subsidy (and p_i < p_i^* is an export tax).

Net government trade-tax revenue r(p) is rebated equally to all citizens as a uniform per-capita lump-sum transfer:

Summing labor income l, specific-factor rents ∑_i=1ⁿ Π_i(p_i), rebated tariff revenue r(p), and consumer surplus s(p) yields aggregate gross social welfare W(p):

Taking the partial derivative of W(p) with respect to p_i and applying Hotelling's Lemma (Π′_i = X_i) and Roy's Identity (s′_i = −d_i) yields the Harberger marginal deadweight-loss expression:

Organized vs. Unorganized Sectors and Gross Lobby Welfare W_i(p)

Each citizen owns one unit of labor and at most one type of sector-specific capital. Let α_i denote the fraction of the voting population that owns the specific capital of sector i. In an exogenous subset of sectors L ⊆ {1, ..., n}, the specific-factor owners overcome Mancur Olson's free-rider problem and organize into an industry lobby i ∈ L. Let I_i ∈ {0, 1} equal 1 if sector i is organized (i ∈ L) and 0 if unorganized (i ∉ L). The fraction of the population represented by organized lobbies is:

Because members of lobby i receive 100% of the specific-factor rent Π_i(p_i) plus a share α_i of labor income, tariff revenue, and consumer surplus, the gross joint welfare W_i(p) of lobby i (prior to subtracting its campaign contribution C_i(p)) is:

Differentiating W_i(p) with respect to its own sector price p_i and with respect to any other sector's price p_j (j ≠ i) reveals the core distributive conflict:

Lobby i wants a higher domestic price p_i for its own output X_i, but opposes high prices p_j on other industries' goods j ≠ i because its members consume those goods as citizens.

Common Agency Equilibrium and Locally Truthful Contribution Schedules

In Stage 1, each organized lobby i ∈ L non-cooperatively offers the incumbent government a differentiable campaign contribution schedule C_i(p) ≥ 0 contingent on the price vector p. In Stage 2, the incumbent government selects p to maximize its political objective function:

where a ≥ 0 is the weight the government places on aggregate gross social welfare W(p) relative to a dollar of campaign contributions.

Applying Bernheim and Whinston's (1986) characterization of menu auctions under common agency (a multi-principal principal–agent problem), an equilibrium price vector p⁰ must maximize the joint payoff W_i(p) − C_i(p) + G(p) of each lobby i ∈ L and the government, yielding the local truthfulness condition: at p⁰, each lobby's marginal contribution matches its marginal willingness to pay:

Summing Marginal Willingness to Pay Across Lobbies and the Government

Substituting ∂C_j(p⁰)/∂p_i = ∂W_j(p⁰)/∂p_i into the government's first-order condition ∂G(p⁰)/∂p_i = 0 for sector i yields:

Substituting the explicit derivatives ∂W_j/∂p_i and ∂W/∂p_i gives:

Solving for the equilibrium price wedge (p_i⁰ − p_i^*) / p_i⁰ = t_i⁰ / (1 + t_i⁰) yields Grossman and Helpman's Proposition 1 (The Protection for Sale / Modified Ramsey Formula):

where:

• t_i⁰ = (p_i⁰ − p_i) / p_i is the equilibrium ad valorem import tariff (if M_i > 0) or export subsidy (if M_i < 0);
• I_i ∈ {0, 1} equals 1 if industry i is politically organized (i ∈ L) and 0 if unorganized (i ∉ L);
• α_L ∈ [0, 1] is the fraction of the population represented by an organized lobby;
• a ≥ 0 is the government's weight on aggregate social welfare relative to campaign contributions;
• z_i = X_i(p_i⁰) / |M_i(p_i⁰)| is the ratio of domestic industry output to imports (or exports)—the inverse import penetration ratio; and
• e_i = − M′_i(p_i⁰) p_i⁰ / |M_i(p_i⁰)| > 0 is the absolute price elasticity of import demand (or export supply).

Economic Intuition and Structural Predictions of the Formula

1. Sign of Protection (I_i − α_L):
   • Organized Sectors (I_i = 1): Provided α_L < 1, 1 − α_L > 0, so every politically organized industry receives positive protection (t_i⁰ > 0: an import tariff for import-competing sectors and an export subsidy for export sectors).
   • Unorganized Sectors (I_i = 0): Provided α_L > 0, −α_L < 0, so every unorganized industry faces negative protection (t_i⁰ < 0: an import subsidy or export tax) because organized lobbies represent a fraction α_L of consumers who prefer cheaper unorganized goods.
2. Universal Symmetrical Lobbying (I_i = 1 for all i and α_L = 1): If 100% of voters belong to organized lobbies (I_i − α_L = 0), every citizen's consumer interest in other sectors cancels out their producer lobbying advantage, restoring universal free trade (t_i⁰ = 0) even as lobbies pay positive campaign contributions to prevent rival lobbies from taxing them.
3. Inverse Import Penetration Ratio (z_i = X_i / M_i) and Ramsey Inverse Elasticity (1 / e_i): For an organized industry (I_i = 1), protection increases with z_i = X_i / M_i because a unit tariff increase raises industry rents by X_i while the deadweight loss is proportional to imports M_i, and decreases with e_i as in Ramsey–Boiteux pricing.

Goldberg and Maggi (1999) and Gawande and Bandyopadhyay (2000)

Pinelopi Koujianou Goldberg and Giovanni Maggi (1999) and Kishore Gawande and Usree Bandyopadhyay (2000) tested the structural equation [ t_i / (1 + t_i) ] e_i = γ (X_i / M_i) + δ I_i (X_i / M_i) + ε_i on U.S. non-tariff barrier coverage ratios and PAC contributions:

• Confirmation of Sign Switching: Both studies confirmed that δ > 0 and γ ≤ 0: higher output-to-import ratios X_i / M_i increase protection in politically organized industries (I_i = 1), resolving the empirical puzzle that unconditional regressions of tariffs on import penetration yielded mixed signs.
• Endogenous Lobby Formation and Firm Heterogeneity: Mitra (1999) incorporated fixed costs of lobby formation (explaining why rising import penetration induces sectors to organize), and Bombardini (2008) showed that within-sector firm-size dispersion increases lobbying effectiveness.

Why does the Protection for Sale model predict higher protection when import penetration is low (X_i / M_i is high)?

Holding lobby organization I_i = 1 fixed, when domestic output X_i is large relative to imports M_i, a 1% tariff increase delivers a large rent gain p_iX_i to domestic specific-capital owners while imposing a small deadweight loss on the small volume of imports M_i. When import penetration rises in raw data, it first induces unorganized industries to pay the fixed cost of forming a lobby (I_i switches from 0 to 1, Mitra 1999); once I_i = 1 is controlled for, protection increases with X_i / M_i.

Why is the model titled "Protection for Sale" rather than "Politicians for Sale"?

In a Bernheim–Whinston menu auction, lobbies do not bribe politicians to abandon social welfare; rather, the government sells trade protection policies along a continuous schedule while placing weight a > 0 on aggregate social welfare W(p). Because multiple lobbies compete against one another, the government extracts political rents from the lobbies—especially when α_L → 1, where lobbies pay contributions merely to defend free trade against rival lobbies.

How does the Protection for Sale model connect to the Bagwell–Staiger GATT/WTO theory?

In their 1995 paper "Trade Wars and Trade Talks," Grossman and Helpman extended the Protection for Sale model to a two-country large open economy where domestic tariffs alter world prices p^*. Each country's non-cooperative Nash tariff equals the domestic Protection for Sale political-economy formula plus the international optimal-tariff markup 1 / ε_i^*. As proved in the Terms-of-Trade Theory of Trade Agreements (Bagwell and Staiger, 1999), a trade agreement strips away the 1 / ε_i^* terms-of-trade markup while leaving the domestic Protection for Sale component intact as the politically optimal tariff.

1. Grossman, G. M., & Helpman, E. (1994). "Protection for Sale." American Economic Review, 84(4), 833–850. https://www.jstor.org/stable/2118033
2. Bernheim, B. D., & Whinston, M. D. (1986). "Menu Auctions, Resource Allocation, and Economic Influence." Quarterly Journal of Economics, 101(1), 1–31. https://doi.org/10.2307/1884639
3. Goldberg, P. K., & Maggi, G. (1999). "Protection for Sale: An Empirical Investigation." American Economic Review, 89(5), 1135–1155. https://doi.org/10.1257/aer.89.5.1135
4. Grossman, G. M., & Helpman, E. (1995). "Trade Wars and Trade Talks." Journal of Political Economy, 103(4), 675–708. https://doi.org/10.1086/262000
5. Gawande, K., & Bandyopadhyay, U. (2000). "Is Protection for Sale? Evidence on the Grossman–Helpman Theory of Endogenous Protection." Review of Economics and Statistics, 82(1), 139–152. https://doi.org/10.1162/003465300558579