Political Economics - Explaining Economic Policy

Transcription

1 Political Economics - Explaining Economic Policy T. Persson and G. Tabellini (Book ; Chapters 1-5) presented by Salvatore Lo Bello Macro Reading Group UC3M November 14, 2012 T. Persson and G. Tabellini (Book ; Chapters Political 1-5) Economics presented- Explaining by Salvatore Economic Lo Bello Policy (Macro Reading November Group UC3M) 14, / 17

2 Introduction The Motivation Economic policies vary greatly across time and place: in the late 1990s total government spending was more than 60% of GDP in Sweden, above 50% in many continental Europe countries, around 35% in Japan, Switzerland, USA. The composition of the spending is also characterized by a great variability across countries (e.g. transfers are high in Europe but low in Latin America). How can we explain the variations in the data? Is there any systematic correlation with other aspects of the economic and social environment? Final goal: explain economic policy in modern democracies, size/form of redistributive programs, extent/type of public good provision, size of government deficits, extent of corruption. We are at the boundary between poltical science and economics. T. Persson and G. Tabellini (Book ; Chapters Political 1-5) Economics presented- Explaining by Salvatore Economic Lo Bello Policy (Macro Reading November Group UC3M) 14, / 17

3 Politics Introduction Motivation of Politicians: opportunistic (office seeking, rent seeking) - standard assumption. partisan (they maximize a social welfare function with disproportionate weights). Timing of Politics: Preelection Politics: electoral promises are binding and enforceable; candidates propose policies to maximize their chances of winning. Voters only like economic policies = median voter theorem = assumptions on the motivation of politicians become unimportant. Voters also like other fixed factors of politicians (ideology) = the motivations of politicians matter. Postelection Politics: electoral promises are not binding or too vague to even matter; voters select the politician, not directly the policy. Winner takes all: one politician free to set the policy. Legislative Bargaining. T. Persson and G. Tabellini (Book ; Chapters Political 1-5) Economics presented- Explaining by Salvatore Economic Lo Bello Policy (Macro Reading November Group UC3M) 14, / 17

4 Preferences Notation and Preferences Heterogeneous agents (characterized by α i specific feature), affected by a policy vector q. W (q, p; α i ) = max c i [U(c i, q, p; α i ) H(c i, q, p; α i ) 0]. The policymaker sets q, respecting the market-determined value of p and some other constraints: G(q, p) 0. The constraint will typically be binding = p = P (q). Therefore, we can define the preferred policy of voter i: q(α i ) = arg max W (q; α i ) q T. Persson and G. Tabellini (Book ; Chapters Political 1-5) Economics presented- Explaining by Salvatore Economic Lo Bello Policy (Macro Reading November Group UC3M) 14, / 17

5 Notation and Preferences Restricting Preferences Arrow (1951) has shown that no general rule enables a democracy to consistently aggregate individual preferences = majority rule does not always generate well-defined equilibrium policies. Definition 1 A Condorcet winner is a policy q that beats any other feasible policy in a pairwise vote. Definition 2 Policy preferences of voter i are single peaked if: If q q q(α i ) (or if q q q(α i )) = W (q ; α i ) W (q ; α i ). Proposition 1 If all the voters have single-peaked preferences over a given ordering of policy alternatives, a Condorcet winner always exists and coincides with the median-ranked bliss point. This equilibrium is also unique. T. Persson and G. Tabellini (Book ; Chapters Political 1-5) Economics presented- Explaining by Salvatore Economic Lo Bello Policy (Macro Reading November Group UC3M) 14, / 17

6 Restricting Preferences Notation and Preferences Definitions 3 The preferences of voters in A satisfy the single-crossing property if: If q > q and α i > α i (or if q < q and α i < α i ), then W (q; α i ) W (q ; α i ) = W (q; α i ) W (q ; α i ) Voters in A have intermediate preferences if: W (q; α i ) = J(q) + K(α i )H(q), where K(α i ) is momotonic in α i. Both these conditions guarantee existence and unicity of the equilibrium. Example - Redistributive Distortionary Taxation w i = c i + V (x i ) c i = (1 q)l i + f, where f ql = ql(q) (gov. budget constraint) 1 α i x i + l i Optimal labor supply: l i = 1 α Vx 1 (1 q) (α i α). W i (q; α i ) = L(q) + V (1 L(q) α) (1 q)(α i α). T. Persson and G. Tabellini (Book ; Chapters Political 1-5) Economics presented- Explaining by Salvatore Economic Lo Bello Policy (Macro Reading November Group UC3M) 14, / 17

7 Notation and Preferences Nonexistence of a Condorcet Winner T. Persson and G. Tabellini (Book ; Chapters Political 1-5) Economics presented- Explaining by Salvatore Economic Lo Bello Policy (Macro Reading November Group UC3M) 14, / 17

8 Electoral Competition A Simple Model of Public Finance Electoral Competition A society inahabited by a continuum of citizens. Government budget constraint: w i = c i + H(g) (1) c i = (1 τ)y i (2) τy = g (3) = W i (g) = (y g) yi y + H(g) = g i = Hg 1 ( yi y ) Normative benchmark: i W i (g)df = W (g) = g = H 1 g (1) T. Persson and G. Tabellini (Book ; Chapters Political 1-5) Economics presented- Explaining by Salvatore Economic Lo Bello Policy (Macro Reading November Group UC3M) 14, / 17

9 Electoral Competition Downsian Electoral Competition 1 Candidates A,B commit to a policy g, in order to maximize the chance of winning p. 2 Elections are held. 3 The elected candidate implements his announced policy. 0 if W m (g A ) < W m (g B ) p A = 1 2 if W m (g A ) = W m (g B ) 1 if W m (g A ) > W m (g B ) Trivially, the equilbrium will be: g m = Hg 1 ( ym y ) = Suboptimality. T. Persson and G. Tabellini (Book ; Chapters Political 1-5) Economics presented- Explaining by Salvatore Economic Lo Bello Policy (Macro Reading November Group UC3M) 14, / 17

10 Electoral Competition Probabilistic Voting Candidates may differ in other dimensions unrelated to the policy (ideology, a second policy dimension in which they cannot make credible commitments). Three groups: R,M,P (y R > y M > y P ). Share of group j is α j, such that j α j = 1. Voter i of group J prefers candidate A if: W J (g A ) > W J (g B ) + σ ij + δ (4) σ ij and δ are distributed as U[ 1 2φ j, 1 2φ j ] and U[ 1 2ψ, 1 2ψ ]. Swing voter of group j: W J (g A ) = W J (g B ) + σ j + δ T. Persson and G. Tabellini (Book ; Chapters Political 1-5) Economics presented- Explaining by Salvatore Economic Lo Bello Policy (Macro Reading November Group14, UC3M) / 17

11 Probabilistic Voting Electoral Competition 1 The two candidates announce their electoral platforms: g A,g B. 2 The actual value of δ is realized and all the uncertainty is resolved. 3 Elections are held. 4 The elected candidate implements his announced policy. Candidate A will maximize the following: π A = ( α j φ j σ j + 1 ) 2φ j J [ p A = P rob π A 1 ] [ ] = ψ α j φ j [W j (g A ) W j (g B )] φ J where φ = J αj φ j is the average density across groups. F OC : α j φ j H g (g) = 1 ( J α j φ j y j = g S = H 1 αj φ j y j ) g y φy J J T. Persson and G. Tabellini (Book ; Chapters Political 1-5) Economics presented- Explaining by Salvatore Economic Lo Bello Policy (Macro Reading November Group14, UC3M) / 17 (5) (6)

12 Probabilistic Voting Electoral Competition Figure : Electorate in a Probabilistic Voting Model Figure : Bliss Points of Different Swing Voters. Persson and G. Tabellini (Book ; Chapters Political 1-5) Economics presented- Explaining by Salvatore Economic Lo Bello Policy (Macro Reading November Group14, UC3M) / 17

13 Partisan Politicians Partisan Politicians Policy Convergence (binding commitments) Two exogenous candidates (L, R), same timing as before. 0 if W m (g L ) < W m (g R ) p L = 1 2 if W m (g L ) = W m (g R ) 1 if W m (g L ) > W m (g R ) Candidate L maximizes: E[W L (g)] = p L W L (g L ) + (1 p L )W L (g R ) = g L = g R = g m Policy Divergence (no binding commitments) Only one credible announcemnt for L: g L = Hg 1 ( yl y ) Candidate L wins if W m (g L ) > W m (g R ) T. Persson and G. Tabellini (Book ; Chapters Political 1-5) Economics presented- Explaining by Salvatore Economic Lo Bello Policy (Macro Reading November Group14, UC3M) / 17

14 Partisan Politicians Partisan Politicians - Endogenous Candidates 1 Any citizen can enter as a candidate at a cost of ɛ. 2 Elections are held. 3 The elected candidate sets the policy g P ; if nobody runs, ḡ is implemented. No policy commitment = g p = H 1 g ( yp y ). Unicity of equlibrium under : W m (g m ) W m (ḡ) ɛ. Otherwise, infinitely many equilibria under: W m (g R ) = W m (g L ) 1 2 [W R (g R ) W R (g L )] ɛ 1 2 [W L (g L ) W L (g R )] ɛ T. Persson and G. Tabellini (Book ; Chapters Political 1-5) Economics presented- Explaining by Salvatore Economic Lo Bello Policy (Macro Reading November Group14, UC3M) / 17

15 Agency Problems Agency Problems Can the voters discipline rent-seeking politicians? Gov. budget constraint: τy = g + r Candidates now maximize E(v P ) = p P (R + γr) Under efficient electoral competition: Voters preferences: W i (g) = (y (g + r)) yi y + H(g) 0 if W m (g A ) < W m (g B ) p A = 1 2 if W m (g A ) = W m (g B ) 1 if W m (g A ) > W m (g B ) ( ) y g A = g B = g m = Hg 1 m y r A = r B = r m = 0 T. Persson and G. Tabellini (Book ; Chapters Political 1-5) Economics presented- Explaining by Salvatore Economic Lo Bello Policy (Macro Reading November Group14, UC3M) / 17

16 Agency Problems Agency Problems Under inefficient electoral competition (probabilistic voting): [E(v A )] g A [E(v A )] r A p A = ψ[w (g A, r A ) W (g A, r B )] (7) = (R + γr A ) p A g A = (R + γr A )ψw g (g A, r A ) = 0 (8) = (R + γr A ) p A r A + p A γ = (R + γr A )ψ (9) Positive rents in equilibrium. p A = ψw r = ψ (10) r A [ 1 = r = max 0, 2ψ R ] γ T. Persson and G. Tabellini (Book ; Chapters Political 1-5) Economics presented- Explaining by Salvatore Economic Lo Bello Policy (Macro Reading November Group14, UC3M) / 17

17 Summary Summary We analyzed two different electoral competition models: Downsian model and probabilistic voting = policy convergence. Introducing partisan politicians it is possible to obtain policy divergence. We analyzed the conflict of interests between voters and rent-seeking politicians. It is possible to obtain positive rents in equilibrium. We abstained from agency problems in postelection politics models and from legislative bargaining. T. Persson and G. Tabellini (Book ; Chapters Political 1-5) Economics presented- Explaining by Salvatore Economic Lo Bello Policy (Macro Reading November Group14, UC3M) / 17