Global Liquidity, House Prices, and the Macroeconomy: Evidence from Advanced and Emerging Economies

Transcription

1 Global Liquidity, House Prices, and the Macroeconomy: Evidence from Advanced and Emerging Economies Ambrogio Cesa-Bianchi 1 Luis F. Cespedes 2 Alessandro Rebucci 3 1 Bank of England & Centre for Macroeconomics 2 Universidad Adolfo Ibanez 3 Johns Hopkins University Carey Business School Housing Markets, Monetary Policy and Macroprudential Policy Bank of Finland (October 215) 1

2 Disclaimer The views expressed in this paper are solely those of the authors and should not be taken to represent those of the Bank of England. Introduction 2

3 Motivation Booms and busts in the non-tradable sector, often fuelled by excessive credit expansion and overvalued exchange rates Surges and sudden reversals in cross-border capital flows Introduction 3

4 Motivation Booms and busts in the non-tradable sector, often fuelled by excessive credit expansion and overvalued exchange rates Surges and sudden reversals in cross-border capital flows Housing and global liquidity Housing: quintessential non-tradable asset/durable good Global liquidity: key determinant of international capital flows Introduction 3

5 What we do & Preview of results New quarterly house price data set for 33 emerging markets from 199 to 212 Introduction 4

6 What we do & Preview of results New quarterly house price data set for 33 emerging markets from 199 to 212 New set of house price stylized facts Introduction 4

7 What we do & Preview of results New quarterly house price data set for 33 emerging markets from 199 to 212 New set of house price stylized facts Identify the impact of a global liquidity shock on house prices, and trace its impact on the macro-economy in both AEs and EMs using a panel VAR Consumption, house prices and exchange rates ( ), current account ( ) House prices and exchange rates play a significant amplification role Introduction 4

8 Literature review Global house price cycle [Andre (21); Hirata et al. (212); Igan and Loungani (212); Claessens et al. (212); Cesa-Bianchi (213)] House prices and capital flows [Laibson and Mollerstrom (21); Favilukis et al. (212); Adamet al. (212); Ferrero (212); Aizenman and Jinjarak (29); Gete (29); Sa et al. (214)] Global liquidity [Landau (213),Rey (213); Bruno and Shin (214); Cerutti et al. (214)] Introduction 5

9 Outline Data & (selection of) stylized facts Panel VAR & Global liquidity shocks Inspecting the transmission mechanism Introduction 6

10 Data Unbalanced panel of 57 time series with varying coverage from 199:Q1 212:Q4 Source: OECD, BIS, Dallas FED international house price databases National central banks, national statistical offices, and academic publications on housing markets Data & (selection of) stylized facts 7

11 Data Unbalanced panel of 57 time series with varying coverage from 199:Q1 212:Q4 Source: OECD, BIS, Dallas FED international house price databases National central banks, national statistical offices, and academic publications on housing markets Value added Additional countries: Argentina, Brazil, Chile, Colombia, Czech Republic, India, Serbia, Taiwan, and Uruguay Historical data: China, Estonia, Hong Kong, Hungary, Indonesia, Lithuania, Malaysia, Philippines, Poland, Slovakia, Slovenia, and Thailand Data & (selection of) stylized facts 7

12 Data Map: Advanced Economies (a) Advanced Economies AUS AUT BEL CAN DNK FIN FRA DEU GRC IRL ITA JPN LUX MLT NLD NZL NOR PRT ESP SWE CHE GBR USA ISL Data & (selection of) stylized facts 8

13 Data Map: Emerging Economies (b) Emerging Economies ARG CHL COL HRV HKG KOR MYS SGP ZAF URY THA BGR IDN ISR PHL TWN SVN SRB UKR EST HUN LTU PER CHN CZE POL BRA IND LVA RUS SVK MEX MAR Data & (selection of) stylized facts 9

14 House price inflation strongly pro-cyclical, leads the monetary policy cycle, some (weak) association with CA and RER in AEs (a) Advanced Economies GDP Consumption CPI Cross border Bank Flows Equity Price.5 Short term Int. Rate.5 Real Eff. Exch. Rate Current Account / GDP Data & (selection of) stylized facts 1

15 Similar patterns in EMs: weaker association with monetary cycle and RER; stronger association with CA (b) Emerging Economies GDP Consumption CPI Cross border Bank Flows Equity Price.5 Short term Int. Rate.5 Real Eff. Exch. Rate Current Account / GDP Data & (selection of) stylized facts 11

16 Global Liquidity: Definition & Data Definition Global liquidity (GL) is a vector of supply factors affecting provision of cross-border credit by global banks Panel VAR & Global liquidity shocks 12

17 Global Liquidity: Definition & Data Definition Global liquidity (GL) is a vector of supply factors affecting provision of cross-border credit by global banks US monetary policy = US Interest rates, US M2 Global banks funding conditions = US TED spread, Leverage, US Yield curve slope Risk appetite and uncertainty = VIX Panel VAR & Global liquidity shocks 12

18 Global Liquidity: Definition & Data Definition Global liquidity (GL) is a vector of supply factors affecting provision of cross-border credit by global banks US monetary policy = US Interest rates, US M2 Global banks funding conditions = US TED spread, Leverage, US Yield curve slope Risk appetite and uncertainty = VIX Data Sum (across all countries) of international cross-border liabilities of country i vis-a-vis the rest of the world Panel VAR & Global liquidity shocks 12

19 Model: Panel VAR Panel VAR: Equation for country i x it = a i + b i t + c i t 2 + F 1i x i,t 1 + F 2i x i,t 2 + u it, where x it = [ GL t C it HP it RIR it RER it (CA/Y ) it ] Sample: 1995Q4-212:Q4 (23 AEs and 27 EMEs) Panel VAR & Global liquidity shocks 13

20 Model: Panel VAR Panel VAR: Equation for country i x it = a i + b i t + c i t 2 + F 1i x i,t 1 + F 2i x i,t 2 + u it, where x it = [ GL t C it HP it RIR it RER it (CA/Y ) it ] Sample: 1995Q4-212:Q4 (23 AEs and 27 EMEs) Mean group estimator = Dynamic panel data models with heterogenous slope coefficients Estimate country by country with OLS Take average of IRFs across countries Avoids potential inconsistency issues (Pesaran and Smith, 1995) Panel VAR & Global liquidity shocks 13

21 Identification: Global Liquidity Shock Challenge: disentangling push versus pull. Identification is achieved in two steps Panel VAR & Global liquidity shocks 14

22 Identification: Global Liquidity Shock Challenge: disentangling push versus pull. Identification is achieved in two steps Aggregation Idiosyncratic pull shocks wash away for large N Panel VAR & Global liquidity shocks 14

23 Identification: Global Liquidity Shock Challenge: disentangling push versus pull. Identification is achieved in two steps Aggregation Idiosyncratic pull shocks wash away for large N External instruments [Stock and Watson, 212; Mertens and Ravn, 213] Use the drivers of GL as instruments Isolate the variation of the GL reduced-form residuals that are due only to supply push factors As instruments are U.S. variables, drop U.S. from sample Panel VAR & Global liquidity shocks 14

24 In AEs, GL shock increases house prices, consumption, and affects external sector. Monetary policy tightened as a response (a) Advanced Economies Cross Border Credit Consumption House Price Real Int. Rate Real Eff. Exch. Rate Current Account / GDP Panel VAR & Global liquidity shocks 15

25 In EMs, effects much larger. Transmission mechanism also possibly different (b) Emerging Economies Cross Border Credit Consumption House Price Real Int. Rate Real Eff. Exch. Rate Current Account / GDP Panel VAR & Global liquidity shocks 16

26 Inspecting the transmission mechanism How can we explain the different response of AEs and EMs? Conjecture Global liquidity shock relaxes borrowing constraints via an increase in the value of collateral Inspecting the transmission mechanism 17

27 Inspecting the transmission mechanism How can we explain the different response of AEs and EMs? Conjecture Global liquidity shock relaxes borrowing constraints via an increase in the value of collateral Through which channels? Consider the following borrowing constraint D t θq t h t House prices expand borrowing capacity through increased value of collateral Inspecting the transmission mechanism 17

28 Inspecting the transmission mechanism How can we explain the different response of AEs and EMs? Conjecture Global liquidity shock relaxes borrowing constraints via an increase in the value of collateral Through which channels? Consider the following borrowing constraint D t θq t h t House prices expand borrowing capacity through increased value of collateral If borrowing is in foreign currency, exch. rate appreciation also increases the value of collateral Inspecting the transmission mechanism 17

29 Inspecting the transmission mechanism How can we explain the different response of AEs and EMs? Conjecture Global liquidity shock relaxes borrowing constraints via an increase in the value of collateral Through which channels? Consider the following borrowing constraint D t θq t h t House prices expand borrowing capacity through increased value of collateral If borrowing is in foreign currency, exch. rate appreciation also increases the value of collateral The larger the LTV ratio (θ) the larger the effect of house prices / exch. rates movements on borrowing capacity Inspecting the transmission mechanism 17

30 Inspecting the transmission mechanism (cont d) In a new paper Housing, Leverage, and Global Liquidity (joint with Andrea Ferrero and Alessandro Rebucci) we explore the transmission mechanism in detail Inspecting the transmission mechanism 18

31 Inspecting the transmission mechanism (cont d) In a new paper Housing, Leverage, and Global Liquidity (joint with Andrea Ferrero and Alessandro Rebucci) we explore the transmission mechanism in detail New empirical evidence Country groups based on cross-sectional information (rather than AEs vs EMEs) Foreign currency borrowing LTV ratios / Mortgage debt over GDP Inspecting the transmission mechanism 18

32 Inspecting the transmission mechanism (cont d) In a new paper Housing, Leverage, and Global Liquidity (joint with Andrea Ferrero and Alessandro Rebucci) we explore the transmission mechanism in detail New empirical evidence Country groups based on cross-sectional information (rather than AEs vs EMEs) Foreign currency borrowing LTV ratios / Mortgage debt over GDP Model Simple DSGE model of international borrowing and lending with financial frictions Inspecting the transmission mechanism 18

33 Share of foreign currency liabilities Ukraine Uruguay Philippines Indonesia Poland Latvia Norway Argentina Lithuania Slovakia Peru India Croatia Denmark Israel Hungary United Kingdom Korea Morocco Colombia Estonia Russia Singapore New Zealand Sweden Thailand Slovenia Mexico Austria Czech Republic Brazil Malaysia Finland Switzerland China Canada Australia Greece Japan Chile Hong Kong South Africa Italy Belgium Germany Ireland Netherlands Portugal Iceland United States SpainFrance Note. Data is from Lane and Shambaugh (24,AER). Inspecting the transmission mechanism 19

34 GL shock: low share of foreign currency liabilities Cross Border Credit Consumption House Price Real Int. Rate Real Eff. Exch. Rate Current Account / GDP Inspecting the transmission mechanism 2

35 GL shock: high share of foreign currency liabilities Cross Border Credit Consumption House Price Real Int. Rate Real Eff. Exch. Rate Current Account / GDP Inspecting the transmission mechanism 21

36 A simple model Model: Based on Justiniano, Primiceri and Tambalotti (214) Two countries, two goods, no aggregate uncertainty Inspecting the transmission mechanism 22

37 A simple model Model: Based on Justiniano, Primiceri and Tambalotti (214) Two countries, two goods, no aggregate uncertainty Home households Want to buy housing Subject to collateral constraint: s td t θq th t Inspecting the transmission mechanism 22

38 A simple model Model: Based on Justiniano, Primiceri and Tambalotti (214) Two countries, two goods, no aggregate uncertainty Home households Want to buy housing Subject to collateral constraint: s td t θq th t Foreign households More patient than Home households: β > β Save via deposits and equity in financial intermediaries Inspecting the transmission mechanism 22

39 A simple model Model: Based on Justiniano, Primiceri and Tambalotti (214) Two countries, two goods, no aggregate uncertainty Home households Want to buy housing Subject to collateral constraint: s td t θq th t Foreign households More patient than Home households: β > β Save via deposits and equity in financial intermediaries Global financial intermediaries Channel funds internationally from lenders to borrowers Subject to leverage constraint (capital requirement) Adjustment costs to change equity position Inspecting the transmission mechanism 22

40 Global Financial Intermediaries Balance sheet in at time t (after borrowers and lenders decisions) Assets Liabilities Loans nd t Deposits (1 n)d t Equity (1 n)e t Inspecting the transmission mechanism 23

41 Global Financial Intermediaries Balance sheet in at time t (after borrowers and lenders decisions) Assets Liabilities Loans nd t Deposits (1 n)d t Equity (1 n)e t Maximize profits subject to leverage constraint (capital requirement) nd t χ(1 n)e t Foreign credit supply shock χ t = χ(1 ρ) + ρχ t 1 + ε χ Inspecting the transmission mechanism 23

42 A foreign credit supply shock Borrowing Consumption House Price Real Int. Rate Real Exch. Rate x Current Account Inspecting the transmission mechanism 24

43 Conclusions Consumption, house prices, and external variables respond strongly to liquidity conditions at the center Inspecting the transmission mechanism 25

44 Conclusions Consumption, house prices, and external variables respond strongly to liquidity conditions at the center Channels of transmission Exchange rate and house prices LTV ratio Hard to disentangle from the data Inspecting the transmission mechanism 25

45 Conclusions Consumption, house prices, and external variables respond strongly to liquidity conditions at the center Channels of transmission Exchange rate and house prices LTV ratio Hard to disentangle from the data DSGE model in line with empirical evidence Inspecting the transmission mechanism 25

46 Conclusions Consumption, house prices, and external variables respond strongly to liquidity conditions at the center Channels of transmission Exchange rate and house prices LTV ratio Hard to disentangle from the data DSGE model in line with empirical evidence Next steps Counterfactuals using the DSGE Estimation (IRF matching) Inspecting the transmission mechanism 25

47 Appendix Appendix 26

48 LTV ratio 1 Netherlands Switzerland Denmark New Zealand United States.9 United Kingdom Australia Singapore Sweden Norway.8 Portugal Germany Ireland Canada Hong Kong.7 Spain Japan Finland Malaysia Belgium.6 Austria Taiwan France South Africa.5 Thailand Korea Israel Italy Chile Greece.4 Latvia Estonia Colombia China Mexico.3 Hungary Croatia Morocco Philippines Slovakia.2 India Czech Republic Poland Brazil Peru.1 Indonesia Bulgaria Argentina Slovenia Russia Appendix Additional results 27

49 GL shock: low LTV ratios Cross Border Credit Consumption House Price Real Int. Rate Real Eff. Exch. Rate Current Account / GDP Appendix Additional results 28

50 GL shock: high LTV ratios Cross Border Credit Consumption House Price Real Int. Rate Real Eff. Exch. Rate Current Account / GDP Appendix Additional results 29

51 GL - Data International cross border claims of BIS reporting banks vis a vis the banking sector Trillions US DOllars Constant prices (28:Q2 US Dollars) Growth rate (year on year) Appendix Global liquidity 3

52 A simple model Model: Based on Justiniano, Primiceri and Tambalotti (214) Two countries, two goods, no aggregate uncertainty Appendix Model 31

53 A simple model Model: Based on Justiniano, Primiceri and Tambalotti (214) Two countries, two goods, no aggregate uncertainty Impatient Home households (i [, n)) Want to buy housing Subject to collateral constraint Appendix Model 31

54 A simple model Model: Based on Justiniano, Primiceri and Tambalotti (214) Two countries, two goods, no aggregate uncertainty Impatient Home households (i [, n)) Want to buy housing Subject to collateral constraint Patient foreign households (i (n, 1]) Save via deposits and equity in financial intermediaries Foreign households own financial intermediaries Appendix Model 31

55 A simple model Model: Based on Justiniano, Primiceri and Tambalotti (214) Two countries, two goods, no aggregate uncertainty Impatient Home households (i [, n)) Want to buy housing Subject to collateral constraint Patient foreign households (i (n, 1]) Save via deposits and equity in financial intermediaries Foreign households own financial intermediaries Global financial intermediaries Channel funds internationally from lenders to borrowers Subject to leverage constraint (capital requirement) Adjustment costs to change equity position Appendix Model 31

56 Goods Market Each country endowed with one good, households consume both (c F t )1 α c t (c Ht) α (c F t ) 1 α α α (1 α) 1 α c t (c Ht )α α α (1 α ) 1 α with α (n, 1] and α [, n) Appendix Model 32

57 Goods Market Each country endowed with one good, households consume both (c F t )1 α c t (c Ht) α (c F t ) 1 α α α (1 α) 1 α c t (c Ht )α α α (1 α ) 1 α with α (n, 1] and α [, n) Price indexes (LOOP holds: P it = Pit for i = {H, F }) P t = (P Ht ) α (P F t ) 1 α Pt = (PHt) α (PF t) 1 α Appendix Model 32

58 Goods Market Each country endowed with one good, households consume both (c F t )1 α c t (c Ht) α (c F t ) 1 α α α (1 α) 1 α c t (c Ht )α α α (1 α ) 1 α with α (n, 1] and α [, n) Price indexes (LOOP holds: P it = Pit for i = {H, F }) P t = (P Ht ) α (P F t ) 1 α Pt = (PHt) α (PF t) 1 α Relative prices and real exchange rate terms of trade (τ t P F t /P Ht ) p Ht P Ht /P t = τ α 1 t p F t = P F t/p t = τ α t s t P t /P t = τ α α t Appendix Model 32

59 Households Home country max {ct,h t,d t} U t = t= βt [u(c t ) + v(h t )] subject to and c t + q t h t s t (d t R t 1 d t 1 ) = p Ht y t + q t h t 1 s t d t θq t h t Credit denominated in units of foreign consumption Appendix Model 33

60 Households Home country max {ct,h t,d t} U t = t= βt [u(c t ) + v(h t )] subject to and c t + q t h t s t (d t R t 1 d t 1 ) = p Ht y t + q t h t 1 s t d t θq t h t Credit denominated in units of foreign consumption Foreign country (β > β) max {c t,d t } U t = t= β t u(c t ) subject to c t + d t + e t = p F t y t + R d t 1d lt 1 + Re t 1e t 1 + Π t Foreign households own financial intermediaries Appendix Model 33

61 Global Financial Intermediaries Balance sheet in at time t (after borrowers and lenders decisions) Assets Liabilities Loans nd t Deposits (1 n)d t Equity (1 n)e t Appendix Model 34

62 Global Financial Intermediaries Balance sheet in at time t (after borrowers and lenders decisions) Assets Liabilities Loans nd t Deposits (1 n)d t Equity (1 n)e t Next period profits for financial intermediaries Π t+1 = R t nd t R d t (1 n)d t [1 + Ψ(e t )]R e t (1 n)e t where Ψ(e t ) η(e t /ē) γ (η > and γ > 1) is equity adjustment cost Appendix Model 34

63 Global Financial Intermediaries Balance sheet in at time t (after borrowers and lenders decisions) Assets Liabilities Loans nd t Deposits (1 n)d t Equity (1 n)e t Next period profits for financial intermediaries Π t+1 = R t nd t R d t (1 n)d t [1 + Ψ(e t )]R e t (1 n)e t where Ψ(e t ) η(e t /ē) γ (η > and γ > 1) is equity adjustment cost Subject to leverage constraint (capital requirement) nd t χ(1 n)e t Appendix Model 34

64 Global Financial Intermediaries Profit maximization yields solution for R t R t = 1 χ [1 + (1 + γ)ψ(e t)]r e t + ( 1 1 ) Rt d χ Assume risk-neutral foreign households ( Rt d = Rt e = 1/β l ) R t = 1 [ β 1 + (1 + γ)ψ(e ] t) χ Equilibrium with binding credit supply constraint (nd t = χ(1 n)e t ) R t = 1 { [ ] γ } η(1 + γ) nd t β 1 + χ χ(1 n)ē Equilibrium with non-binding credit supply constraint not interesting R t = R d t = 1/β e t = Appendix Model 35

65 Global Credit Market Equilibrium Assume fixed supply of housing h t = h = 1 Domestic households (also risk-neutral) (1 θµ t )q t = mrs + βq t+1 1 µ t = βr t s t+1 /s t µ t and s t d t θq t Foreign households R d t = R e t = 1/β Financial intermediaries (binding credit constraint: nd t = χ(1 n)e t ) R t = 1 { [ ] γ } η(1 + γ) nd t β 1 + χ χ(1 n)ē Appendix Model 36

66 Goods Market Equilibrium Resource constraint ny t = nc Ht + (1 n)c Ht (1 n)y t = nc F t + (1 n)c F t Replace consumption demands ny t = τ 1 α t [αnc t + s t α (1 n)c t ] (1 n)yt = τ α t [(1 α)nc t + s t (1 α )(1 n)c t ] Borrower s budget constraint c t = τ α 1 t y t + s t (d t R t 1 d t 1 ) Given d t, three equations in c t, c t and τ t (s t = τ α α t ) Appendix Model 37