The joy of flying: efficient PPP airport concession contracts Eduardo Engel, Ronald Fischer and Alexander Galetovic U. de Chile, U de Chile, U. de los Andes Three organizational forms to provide infrastructure • Public/traditional – Government plans and owns the infrastructure – Government invests, operates and runs the infrastructure, many times subcontracting tasks, mostly in unbundled fashion – Direct link with budget • Public-private partnership (PPP) – Government plans and owns the infrastructure – Private firm invests, builds, operates, runs and then transfers the infrastructure back to the government (bundled contract) – Direct (if opaque) link with budget • Privatization – Private firm plans and owns the infrastructure – Private firm invests, builds, operates and runs the infrastructure – No link with budget Many airports are privately run • • • • PPPs PPIAF database: 141 airports in low & middle income countries are PPPs; 110 involve investment & operation 20-50 years; 30 on average Greece, Zagreb, Belgrade, Mynamar, Cuzco San Juan, (Kennedy), (La Guardia), Madrid, Sydney, Kansai, Osaka, 22 smaller airports in France, … Privatization • Europe: Heathrow, Rome, Vienna, Copenhagen, Manchester, …. Variable, exogenous demand Non-aviation revenues • Airports have two revenue sources – Aviation revenue(e.g. passenger charges, landing fees, terminal rentals) – Non-aviation revenue (e.g. retail concessions, car parking, real estate rents) • Non-aviation revenue is a function of – The exogenous demand for travel – Effort, investment & effort of the operator • Non-aviation revenue is important Global airport revenues & costs (2012) The problem Design a PPP contract that: (i) Reduces or eliminates costly exogenous, (aviation-related) demand risk (ii) Provides incentives to exert efficient effort to increase non-aviation revenues (iii) Can be awarded in a competitive auction → Exploit correlation between aviation and nonaviation revenue The model • Infrastructure – Costs I – Does not depreciate • Exogenous demand risk (aviation revenue) – PV of user fees: v, p.d.f. f(.) – Corresponds to willingness to pay – I vmin v vmax The model II • Endogenous risk (non-aviation revenue) – Concessionaire exerts non-contractible “effort” (investment) e ≥ 0 before operation – With probability p(e) generates value 𝜃v – With probability 1 − p(e) generates no value – p(0) ≥ 0; p(e) < 1; p’ > 0; p” < 0; p’(∞) = 0 • Risk-neutral planner • Many risk averse firms – VNM utility U(y,e) = u(y) ─ ke – Outside option U(0,0) = u(0) The model III Principal chooses payment schedules which depend on v and failure (f) or success (s) of nonaviation project: {Rf(v);Rs(v)}, with 0 R f (v) v , 0 Rs (v) v v ; and effort e Principal’s problem max (1 p(e)) [v R f (v)] f (v)dv p(e) [(1 )v Rs (v)] f (v)dv s.t. u(0) (1 p(e)) u(R f (v) I) f (v)dv p(e) u(Rs (v) I)] f (v)dv ke e argmax e (1 p(e)) u(R f (v) I) f (v)dv p(e) u(Rs (v) I)] f (v)dv ke 0 R f (v) v 0 Rs (v) v v e 0. Result 1: concessionaire bears no exogenous risk Take any {Rf(v);Rs(v)} that solves the principal’s problem and replace it by {Rf; Rs} such that u(R f I) u(R f (v) I) f (v)dv , u(Rs I) u(Rs (v) I)] f (v)dv. Then both the PC and the ICC hold by construction but R f R f (v) f (v)dv , Rs Rs (v) f (v)dv , Reformulated principal’s problem max (1 p(e))[v R f ] p(e)[(1 )v Rs ] s.t. u(0) (1 p(e))u(R f I) p(e)u(Rs I) ke k p(e) u(Rs I) u(R f I) e0 Result 2: some effort increases welfare Let (e) p(e) e / p(e) . If I 1 (e) v there exists a contract of the form Rs (1 )R f with e* 0 and (0,1] such that the principal improves upon the contract Rs R f I Result 3: cross-subsidy from non-aviation to aviation business Rs (1 )R f I R f I Rs (1 )R f R f Rs (1 )R f R f I Losses in both states Profits in both states Result 4: smaller Rf , more effort Define e(Rf) via p(e) u((1 )R f I) u(R f I) k Because p’’< 0, to have e(Rf) decreasing, we need R f u((1 )R f I) u(R f I) decreasing in Rf Result 4: smaller Rf , more effort Defining J(R , ; ) u((1 )R f I) , the condition for e(Rf ) decreasing in Rf holds for all αθ > 0 if 2 J (R f , ) 0 R f Sufficient condition: ((1 )R f I) . 1 Economics: when condition holds and aviation revenues are smaller, larger marginal return to effort Result 5: monotonicity Now the principal’s objective function V (R f ) 1 p(R f ))[v R f ] p(e)[(1 )v Rs ] is decreasing in Rf; and the concessionaire’s PC U(R f ) (1 p(R f ))u(R f I) p(R f )u((1 )R f I) ke(R f ) is increasing in Rf, R f (I / (1 ), I) Result 5: monotonicity Result 6: an PVR auction bidding on Rf implements the contract • Given α, firms bid on Rf • Lowest bid wins • Concession lasts until winning bid is collected in aviation revenues • Concessionaire collects R f in non-aviation revenue if project is successful * • Competition forces to bid R f • Works even if principal does not observe ancillary revenue and 1 “Result” 7: optimal risk sharing (α) Summary • Efficient PPP contract implementable with a PVR auction on aviation revenue only • R f I :concessionaire loses money in aviation business • Concessionaire earns money in non-aviation business; cross subsidy from non-aviation business • Concessionaire is shielded against exogenous demand risk in both businesses • Fixed-term concession is not optimal • Competition takes care of rent-extraction • Works even if principal does not observe ancillary revenue (value) Thank you