Financial Liberalization, Debt Mismatch, Allocative Efficiency and - PowerPoint PPT Presentation

Financial Liberalization, Debt Mismatch, Allocative Efficiency and Growth Romain Ranciere and Aaron Tornell May 11, 2015

Can one make the case for Financial liberalization? ◮ Our answer: a qualified yes ◮ Provided regulation imposes limits on the type of issuable liabilities

Financial liberalization may enhance growth and consumption possibilities because it improves allocative efficiency: ◮ By allowing for new financing instruments and the undertaking of risk, liberalization relaxes financing constraints. ◮ Sectors more dependent on external finance can invest more and grow faster. ◮ The rest of the economy benefits from this relaxation of the bottleneck via input–output linkages.

However, ◮ The use of new financial instruments → a riskless economy is transformed into one with systemic-risk → ↑ Incidence of crises → Bailout costs ◮ ↑↓ Consumption opportunities ◮ We derive a condition for gains from gowth that we bring to the data

Model ◮ Combine endogenous growth model with Schneider-Tornell (2004) ◮ Two-sectors: ◮ Input (N) sector ◮ Final goods (T) sector p n ◮ T-good is numeraire → p t = t p T t ◮ N-sector uses its own goods as capital ◮ φ : share of N-output commanded by the N-sector for investment. ◮ φ determines production efficiency and GDP growth.

Agents: ◮ Risk-neutral investors, opportunity cost 1 + r ◮ Workers (T-sector): supply inelastically l T t = 1 , wage v T t ◮ Entrepreneurs (N-sector): supply inelastically l t = 1 , wage v t ◮ OLGs, linear preferences over consumption of T-goods c t + 1 1+ r c t +1

T-sector ◮ Produce T-goods using N-inputs y t = d α t ( l T t ) 1 − α , α ∈ (0 , 1) . ◮ Representative T-firm maximizes profits taking as given the price of N-goods ( p t ) and standard labor wage ( v T t ) � � y t − p t d t − v T t l T max t d t ,l T t

N-sector ◮ Produce N-goods using entrepreneurial labor ( l t ) , and capital ( k t ) . 1 − β , q t = Θ t k β 1 − β , t l t Θ t =: θk t k t = I t − 1 ◮ Budget constraint p t I t + s t ≤ w t + B t , w t = v t . ◮ Can issue two types of one-period bonds ◮ N-bonds promise to repay in N-goods. ◮ T-bonds promise to repay in T-goods. ◮ Profits π ( p t +1 ) = p t +1 q t +1 +(1+ r ) s t − v t +1 l t +1 − (1+ ρ t ) b t − p t +1 (1+ ρ n t ) b n t .

Production Efficiency ◮ Central Planner allocates supply of inputs ( q t ) to final goods production ( d t = [1 − φ t ] q t ) and to input production ( I t = φ t q t ) . � ∞ � ∞ W po = δ t [ c e δ t [ c t + c e max t + c t ] , s.t. t − y t ] ≤ 0 , { c t ,c e t ,φ t } ∞ t =0 t =0 t =0 y t = [1 − φ t ] α q α q t +1 = θφ t q t . t ,

Optimality → maximizes PV of final goods (T-)production ( � ∞ t =0 δ t y t ) ◮ ↑ φ today → ↓ today’s T-output by α (1 − φ ) α − 1 q α t ∂φ → ↑ tomorrow’s N-output by θq t ∂φ → ↑ tomorrow’s T-output by α [(1 − φ ) θφq t ] α − 1 θq t ∂φ → Intertemporal rate of transformation M = α [(1 − φ ) θφq t ] α − 1 θq t = θ α φ α − 1 . α (1 − φ ) α − 1 q α t ◮ Set M = δ − 1 φ cp = ( θ α δ ) 1 1 − α , if δ < θ − α .

Imperfections Contract Enforceability Problems . If at time t the entrepreneur incurs a non-pecuniary cost H [ w t + B t ] , then at t + 1 she will be able to divert all the returns provided the firm is solvent (i.e., π ( p t +1 ) ≥ 0) . Systemic Bailout Guarantees . If a majority of firms become insolvent, then a bailout agency pays lenders the outstanding liabilities of each non-diverting firm that defaults. Bankruptcy Costs. If a firm is insolvent ( π ( p t +1 ) < 0) a share 1 − µ w of its revenues is lost in bankruptcy procedures. The remainder is paid as wages to the young entrepreneurs.

Regulatory Regimes Financial Repression . Can issue only one-period standard bonds with repayment indexed to the price of N-goods that it produces. Financial Liberalization . Can issue one-period standard bonds with repayments denominated in N- or T-goods. Anything Goes . Can also issue option-like catastrophe bonds.

Symmetric Equilibrium Given prices, N-sector firms and creditors set ( I t , s t , b t , b n t , ρ t , ρ n t ) ; the T-sector demand for N-input d t maximizes T-firms’ profits; factor markets clear; and the market for intermediate goods clears d t ( p t ) + I t ( p t , p t +1 , p t +1 , χ t +1 ) = q t ( I t − 1 ) . � � 1 p t +1 with probability χ t +1 p t +1 = χ t +1 = p t +1 with probability 1 − χ t +1 u ∈ (0 , 1) .

Allocation under Financial Repression There exists an SSE if and only if H ∈ (0 , 1) , β ∈ ( H, 1) and the input � � 1 α � � 1 − α α . sector productivity θ > θ s ≡ 1 1 − β βδ 1 − H ◮ Debt is hedged and crises never occur ( χ t +1 = 1) . ◮ Input sector debt H b n t = 1 − H w t ◮ Investment Share φ s = 1 − β I t = φ s q t , 1 − H .

Bottleneck: ◮ Under financial repression the investment share is below the Central Planner’s optimum: φ s < φ cp � 1 � 1 ◮ Why? φ s < φ cp can be rewritten as θ > α ( φ s ) 1 − α ≡ θ ′ , α δ � � 1 α � 1 � 1 ◮ An equilibrium exists only if θ > θ s ≡ 1 − α α ( φ s ) α . 1 β δ ◮ Since β ∈ (0 , 1) → θ s > θ ′ .

Allocation under Financial Liberalization ◮ Systemic risk: a sunspot can induce a sharp fall in the input price that bankrupts all input sector firms and generates a systemic crisis, during which creditors are bailed out. ◮ There exists an RSE for any crisis’ financial distress costs l d ∈ (0 , 1) if and only if � H � H ∈ (0 , 1) , u ∈ ( H, 1) , β ∈ u , 1 , θ ∈ ( θ, θ ) H/u ◮ Debt is risky: b t = 1 − H/u w t ◮ Input sector’s investment ( I t = φ t q t ) ( τ i denotes a crisis time): � � 1 − u φ l ≡ 1 − β if t � = τ i ; if t � = τ i ; 1 − Hu − 1 χ t +1 = φ t = φ c ≡ µ w 1 if t = τ i ; if t = τ i . 1 − H

= θφ S q ( p ) q − − t t t 1 t 1 1   α − α 1 1 D =   (N-Firms are Solvent) q ( p )   A t t − φ l  p  1 p t t B p t 1  α  1 − α 1 =   D (N-Firms are Bankrupt) q ( p )   t − φ c  p  1 t Figure: Market Equilibrium for Input 18

Bottleneck II: ◮ Under financial Liberalization the investment share is below the Central Planner’s optimum � 1 � 1 α � φ l � 1 − α ◮ φ l < φ cp ⇔ θ > ′′ . α ≡ θ δ ◮ A risky equilibrium exists only if θ > θ. ◮ Can show that θ > θ ′′ for all ( H, u, β, δ ) for which an RSE exists.

GDP Growth gdp t = p t I t + y t Equilibrium N-sector investment, T-output, and prices: y t = [(1 − φ t ) q t ] α , p t = α [(1 − φ t ) q t ] α − 1 . I t = φ t q t , Substituting Z ( φ t ) ≡ 1 − (1 − α ) φ t gdp t = q α t Z ( φ t ) , (1 − φ t ) 1 − α . ◮ Repressed Economy � � α gdp t 1 + γ s ≡ = ( θφ s ) α . θ 1 − β = 1 − H gdp t − 1

Liberalized economy ◮ Tranquil times � � α � θφ l � α . gdp t 1 − β 1 + γ l ≡ = θ = 1 − Hu − 1 gdp t − 1 ◮ Crises can occur. ◮ In equilibrium, 2 crises cannot occur consecutively → average growth in crisis episode �� � 1 / 2 � � 1 / 2 � θ ( φ l φ c ) 1 / 2 � α θφ l � α Z ( φ c ) ( θφ c ) α Z ( φ l ) 1+ γ cr = = . Z ( φ l ) Z ( φ c ) ◮ Loss in GDP growth stems only from the fall in the N-sector’s average investment share ( φ l φ c ) 1 / 2 .

◮ log( gdp t ) − log( gdp t − 1 ) follows a 3-state Markov chain:  � ( θφ l ) α �    log � � 1 − u 0 u   ( θφ l ) α Z ( φ c )  log    . Γ = T = 0 0 1  , Z ( φ l )  � � ( θφ c ) α Z ( φ l ) 1 − u 0 u log Z ( φ c ) ◮ The mean long-run GDP growth rate u u 1 1 − u E (1 + γ r ) = (1 + γ l ) 2 − u (1 + γ cr ) 1 − 2 − u = θ α ( φ l ) 2 − u α ( φ c ) 2 − u α

Growth Enhancing Liberalization E ( γ r ) > γ s ⇔ log( φ l ) − log( φ s ) > [1 − u ] [log(1 − β ) − log( µ w )] where φ c ≡ µ w µ w 1 − β φ s . 1 − H = ◮ Liberalization Enhances Long-run mean GDP growth iff ◮ Benefits of higher leverage and investment in tranquil times ( φ l > φ s ) compensate for the ◮ Shortfall in credit and investment in crisis times ( µ w < 1 − β ) × frequency of crisis (1 − u ) . Let u ↑ 1 → gains for all admissible 1 − u

Panel (a) : Financial Distress Costs and Liberalization Gains Panel (b) : Crisis Probability and Liberalization Gains 4.5 4.5 Financial Repression Financial Repression Financial Liberalization (ld=0.24) Financial Liberalization (crisis proba=2.5%) Financial Liberalization (ld=0.42) 4 4 Financial Liberalization (crisis proba=5%) Financial Liberalization (ld=0.57) Financial Liberalization (crisis proba=7.5%) Financial Liberalization (ld=0.766) 3.5 3.5 3 3 log(GDP) log(GDP) 2.5 2.5 2 2 1.5 1.5 1 1 0.5 0.5 0 10 20 30 40 50 60 70 80 0 10 20 30 40 50 60 70 80 TIME TIME Figure: Growth Gains from Liberalization 33