Winter is Possibly not Coming: Mitigating Financial Instability in an Agent-Based Model with Interbank Market*

Transcription

1 Winter is Possibly not Coming: Mitigating Financial Instability in an Agent-Based Model with Interbank Market* Lilit Popoyan 1, Mauro Napoletano 2,1, Andrea Roventini 1,2 1 Scuola Superiore Sant Anna (Pisa, Italy), 2 OFCE and SKEMA Business School, Sophia-Antipolis (Nice, France) FMM 2017 (Berlin, Germany) 11 November, 2017 * This paper has been supported by European Union s Horizon 2020 grant No Project Dolfins

2 Financial Crisis Effects Excessive Credit Growth United States Euro Area United States Euro Area Figure: Real GDP (left) and credit growth (right) in US and Euro area,

3 Financial Crisis Effects Excessive Credit Growth but not Only!

4 Motivation Macro-prudential policy is linked to monetary policy which alone is too blunt a tool for addressing key systemic risks (Blanchard 2013). Interactions mechanisms between macro-prudential and monetary policy are still unclear (Gambacorta & Murcia 2016, Da Silva 2016). Macro-prudential toolkit is not well explored in terms of its efficiency to contain financial and macroeconomic vulnerabilities: the efficiency of liquidity requirements is highly debated (Bench & Keister 2012, Hessou 2016, Le et al. 2017), a conflict of interest between the liquidity and capital requirements is possible (Elliot&Lindblom 2016, Diamond&Kashyap 2016,Haldane 2015). Potential of corridor width as a macro-prudential instrument (Bulut 2015, Kara 2016)

5 Research Questions What is the role played by monetary policy and prudential regulation in generating market freezes? How do different monetary policy rules and corridor regime interact with macro-prudential regulation under liquidity stress? What are the differences between the Basel II vs. Basel III framework in their capabilities to achieve financial stability when money market rate is defined in corridor system? Which mix of macro-prudential and monetary policy under different corridor regimes can be the best in terms of macrodynamics effects? How do changes in corridor window affect the efficiency of macro-prudential tools and market freeze?

6 Recent Literature Table: Studies on the interaction of macrop-rudential/monetary policy and impact of macro-prudential tools Author(s) of Framework Tool analyzed Impact Interaction Gelain & Ilbas (2017) DSGE CAR + COMP Popoyan et al. (2017) ABM CAR+CCB/LCR/LR + COMP Zhang & Zoli (2016) Empiric LTV + Brzoza-Brzezini et al. (2015) DSGE LTV + COMP Cussen et al. (2015) Empiric LTV, CAR + Krug (2015) ABM CAR+CCB/LR + COMP Aiyar et al. (2014) Empiric CAR - Alpanda et al. (2014) DSGE CAR + COMP Brunnermeier & Sannikov (2014) DSGE LR/CAR - He&Krishnamurthy (2014) DSGE LR/CAR + Quint& Rabanal (2014) DSGE CAR +/- COMP Suh (2014) DSGE CAR - COMP Ozkan&Unsal (2014) DSGE CAR - COMP Angeloni & Faia (2013) DSGE CAR + COMP Benigno et al. (2013) DSGE CAR - Bianchi & Mendoza (2013) DSGE LR - Smets (2013) CC CAR+CCB/LR/LCR + COMP Boissay et al.(2013) DSGE LCR + Lim et al.(2013) Empiric CAR - Angelini et al.(2012) DSGE CAR+CCB/LCR/LR + COMP Adrian& Boyarchenko (2012) DSGE LR + Cincotti et al. 2012) ABM CCB +

7 Our Contribution An analysis of the impact of both Basel II and Basel III s main components on financial vulnerability and macrodynamic under different corridor regimes. An analysis of the interactions between monetary policy and prudential regulation under different interest rate corridor scenarios. The impact of different combination of prudential regulation and monetary policy rules on the macro performance and financial stability of the economy. An analysis of the emerging market freeze, contribution of liquidity requirement and the consequences.

8 Main Results The model is endogenously generating market freeze dynamics. The width and symmetry of interest rate corridor matters form financial stability point of view. Tinbergen principle macro-prudential and monetary policies are not substitutes to reach financial stability. Basel III performs better than Basel II, and it has a stabilizing effect on financial side of economy. LCR is not as efficient to curb with a liquidity crisis: procyclical nature of LCR.

9 Structure of the Model with Interbank Market Extension of Popoyan, L., Napoletano, M. & Roventini, A. 2017

10 Conceptual Framework of the Model Model build on Ashraf, Q., Gershman, B. & Howitt, P and 2017 fixed number N of people with (i,j) type, a fixed number n of different durable goods, n different types of firms, n types of labor, fixed number M of bank sectors, government conducts fiscal policy, central bank conducts the monetary and macro-prudential policy.

11 Stock-Flow Consistent Structure of the Model Table: The balance-sheet matrix of the model Consumer Firm Bank Gov CB Σ Deposit +D c +D s (D c + D s ), +D(R) -D(R) 0 Loans A s +A s, L cb +L cb 0 Loan-IB +IB b, IB b 0 Bond +B b -B +B cb 0 Inventory +I, SC b +SC b +I HPM +H c +H s H cb 0 Balance E c E s E b +GD I Σ

12 Prudential Policy (Basel II and Basel III) Basel II (minimum capital requirement): ɛ 2 i (t) = Total Capital i(t) RWA i (t) = E(T 1) i (t) + E(T 2) i (t) ψ IB IB i (t) + ψ L L s i (t) + ψ SC SC i (t) ɛ2 Basel III: Global capital framework (static capital requirement + counter-cyclical capital buffer, leverage requirement), Global liquidity standard (liquidity coverage ratio).

13 Macro-Prudential Policy Capital Standards Minimum Capital Requirement (CAR3): ɛ 3 i (t) = Tier1 i(t) RWA i (t) = E(T 1) i (t) ψ IB IB i (t) + ψ L L s i (t) + ψ SC SC i (t) ɛ3 Countercyclical Capital Buffer (CCB): 0, if G t < J (G κ = CCB m,t = t J) (H J) 0.025, if J G t H 2.5, if G t > H. Leverage requirement: v i (t) = E i(t) TA i (t) v

14 Macro-Prudential Rules Liquidity Requirement Liquidity Coverage Ratio (LCR): Demand for Liquidity: l i (t) = HQLA i(t) NCOF i (t) l, L d i (t) = l NCOF i(t) HQLA i (t), if LCR < l and zero otherwise. Supply of Liquidity: M j (t) = 1 l HQLA j(t) NCOF j (t), if LCR > l and zero otherwise.

15 Liquidity Trading

16 Central Bank and Monetary Policy Central bank sets nominal interest rate according to: a classical Taylor rule (TR π,y ): ln(1 + i t) = max{ln(1 + i t ) + ϕ π(ln(1 + π t) ln(1 + π t )) + ϕ y (y t y t ), 0} three-mandate Taylor rule (TR π,y,c): ln(1+i t) = max{ln(1+it )+ϕ π(ln(1+π t) ln(1+πt ))+ϕ y (y t yt )+ln( Ct ) ϕc, 0} Ct 1 Central bank defines the interest rate corridor: Baseline corridor: [i t 0.7%, i t + 0.7%], Narrow corridor: [i t 0.5%, i t + 0.5%], Wide corridor: [i t 1%, i t + 1%], Asymmetric corridor: [i t 0.2%, i t + 0.7].

17 Interbank Rate and Central Bank Interest Rate Corridor Model similar to Poole 1968, Whitesell 2006, Ennis & Weinberg 2007 Interest Rate Θ=0 Θ<0 Θ>0 i * +ϕ Demand for IB loans i * Interest corridor i * - ϕ Reserve Balance

18 Interbank Rates Excess supply of reserve of exposures: Θ = n M j (t) j=1 n L d i (t) i=1 Interbank Rates: i + ε i (t), if Θ = 0 i i,t (i, Θ, ε i ) == icorr IB [i,i +φ] + ε i (t), if Θ < 0 icorr IB [i φ,i ] + ε i (t), if Θ > 0 where ε i (t) is a risk premium that depends on the debt-to-equity ratio of the bank i

19 Rate (%) Model Generated Time Series Output, Inflation and Real Interest Rate Inflation Real Interest Rate Output Year 1750

20 U.S. Data vs. Median Model-Generated Outcomes Variable U.S.data Model Inflation Inflation volatility Unemployment Real interest rate Bank failure rate Output gap volatility

21 Cross-Correlation Structure of Output and Other Macro Variables Variable St. Dev t-4 t-3 t-2 t-1 t t+1 t+2 t+3 t+4 GDP Unemployment Inventories Credit Consumption Inflation

22 Phillips Curve Figure: Phillips curve Y = X Inflation Unemployment Note: Inflation over Unemployment (with a line of least squares, R 2 = , p < 0.01, statistical significance at the 5 percent level (**).

23 Okun s Law Figure: Okun s law 10 8 Y = X Percentage change in output Change in unemployment Note: Percentage change in GDP over the percentage change in Unemployment (with a line of least squares, , p < 0.01, statistical significance at the 1 percent level (***).

24 The Mechanics of Expansions and Recessions Credit, Available funds, Failures of Banks and Firm Bankruptcies with Output 1 1 0,5 0, ,5-0,5 Output - Output Output - Credit Lag Output - Output Output - Available Funds Lag 1 1 0,5 0, ,5-0,5 Output - Output Output - Bank failures Lag Output - Output Output - Firm bankruptcies Lag

25 Rate (%) 11/1 12/1 13/01 14/01 15/01 16/01 17/01 18/01 19/01 20/01 21/01 22/01 23/01 24/01 25/01 26/01 27/01 28/01 29/01 30/01 31/01 32/01 33/01 34/01 35/01 36/01 37/01 38/01 39/01 40/01 41/01 42/01 43/01 44/01 45/01 46/01 47/01 48/01 49/01 50/01 51/01 52/01 53/01 54/01 55/01 56/01 57/01 58/01 59/01 60/01 Interbank Market Freezes Figure: Interbank rate dynamics through 50 years 10 9 Policy Rate Interbank Rate Year/Month

26 42/1/1 42/1/2 42/1/3 42/1/4 42/2/1 42/2/2 42/2/3 42/2/4 42/3/1 42/3/2 42/3/3 42/3/4 42/4/1 42/4/2 42/4/3 42/4/4 42/5/1 42/5/2 42/5/3 42/5/4 42/6/1 42/6/2 42/6/3 42/6/4 42/7/1 42/7/2 42/7/3 42/7/4 42/8/1 42/8/2 42/8/3 42/8/4 42/9/1 42/9/2 42/9/3 42/9/4 42/10/1 42/10/2 42/10/3 42/10/4 42/11/1 42/11/2 42/11/3 42/11/4 42/12/1 42/12/2 42/12/3 42/12/4 The Anatomy of Market Freezes Figure: Interbank rate dynamics and excess supply in 42nd year Policy rate Lending Facility Deposit Facility Interbank Rate Q - Supply of Liquidity - Demand of Liquidity Rate (%) Year/Month/Week

27 The Impact of Macro-Prudential Regulation Figure: Radar plot comparison of performance in deferent regulatory tools under the classical Taylor rule Output Gap Bank Failure Rate Unemployment Inflation Likelihood of economic crisis Output gap volatility Basel II Basel III CAR3+CCB+LCR CAR3+CCB CAR3 LCR

28 The Impact of Different Types of Monetary Policies LCR LCR CAR3 CAR3 CAR3+CCB CAR3+CCB CAR3+CCB+LCR CAR3+CCB+LCR Basel III Basel III Basel II Basel II Leaning-against-the-wind TR Classical TR Leaning-against-the-wind TR Classical TR (a) Output gap (b) Unemployment

29 Cyclicality of Macro-Prudential Instruments Output - Output Output - Credit Output - Actual CAR Lag (c) CC Output-actual CAR -0.5 Output - Output Output - Credit Output - Actual CAR+CCB Lag (d) CC Output-Actual CAR+CCB Output - Output Output - Credit Output - Actual LCR Lag (e) CC Output-Actual LCR -0.5 Output - Output Output - Credit Output - Leverage Lag (f) CC Output-Leverage

30 Impact of Different Corridor Regime LCR LCR CAR3 CAR3 CAR3+CCB CAR3+CCB CAR3+CCB+LCR CAR3+CCB+LCR Basel III Basel III Asymmetric Corridor = 0.7 Corridor = 0.5 Corridor = 1 (g) Output gap Asymmetric Corridor = 0.7 Corridor = 0.5 Corridor = 1 (h) Unemployment

31 The Impact of the Interest Rate Corridor Table: Volatility of Economy in Different Corridor Regimes: Normalized values of cross-mc average standard deviation in Money Market Interest, Output Gap and Inflation Baseline Wide Narrow Asymmetric Money market interest (0.1674) (0.0901) (0.0875) Output gap (0.0594) (0.0688) (0.6421) Inflation (0.0784) (0.1014) (0.0946) a Note: (***) significant at 1% level; (**) significant at 5% level; (*) significant at 10% level.

32 Financial Vulnerability Across Prudential Instruments Leverage Leverage GDP Volatility Maturity mismatch GDP Volatility Maturity mismatch Credit-to-GDP gap Size Credit-to-GDP gap Size Concentration Concentration Basel II Basel III CCB+CAR+LCR LCR CCB+CAR (i) Classical Taylor rule Basel II Basel III CCB+CAR+LCR LCR CCB+CAR (j) LAW Taylor rule

33 Conclusing Remarks The model endogenously generates market freezes. The width and symmetry of interest rate corridor matters for financial stability point of view. Macro-prudential and monetary policies are not substitutes (Tinbergen principle). Basel III performs better than Basel II, and it has a stabilizing effect on financial side of economy. LCR is not as efficient to curb with a liquidity crisis: pro-cyclical nature of LCR.

34 Thank You For Your Attention!!! Contact