Liquidity Co-movement Between Financial Institutions and Real Estate Firms: Evidence From China. Sheng Huang

Transcription

1 Liquidity Co-movement Between Financial Institutions and Real Estate Firms: Evidence From China Sheng Huang

2 Background China became the second largest economy in the world in between 2004 to 2014: The total assets of China s banking are fourfold of its domestic GDP, and equivalent to 40% of the global GDP (The Economist, May 2016). China s banking assets increased 448%; The fixed investment in real estate sector in China grew 709%, and the housing price grew 195%. The real estate sector of China contributes roughly 13% of its GDP in average; Chinese commercial banks holds 58% of real estate credit, and the real estate sector receives about 20% of bank loans; Data from the National Bureau of Statistics of China and the People s Bank of China.

3 Background In 2004, five year lending rate increased the first Mean China's Business Environment of Financial Institutions year (mean) Shibor (mean) reserveratioavg (mean) lending rate Libor time in nine years from 5.76% to 6.12%; Bank deposit reserve requirement ratio was raised from 6% in 2000 to 18.16% in 2015; Housing down payment was increased from 20% to 30%, and 5.5 % sales tax has been introduced for resale properties within two years of occupancy. From 2008 to 2009, five year lending rate decreased from 7.83% to 5.94%; deposit reserve requirement ratio declined from 16.3% to 14.1%.

4 Hypothesis A bi-directional liquidity loop exists between financial institutions and real estate firms; Credit risks of real estate sector induce liquidity shortfalls of financial institutions, which in turn curtails liquidity provisions from the financial sector. Liquidity co-movement is stronger during episodes of market distress;

5 Motivation Commonality in liquidity may cause systematic liquidity risks(chordia, et al., 2000; Nikolaou, 2009); Liquidity co-moves across markets (Brockman and Chung, 2009; Phylaktis, 2015; Chordia, et al., 2000), and Firm net cash flow is economically correlated with its market liquidity (Gopalan and Kadan, 2012). The co-movement leads to liquidity risk transmission (Brockman and Chung, 2009; Karolyi and Lee, 2012). The volatility transmission and capital switching between real estate firms and financial institutions have been widely examined, especially at the post subprime market crisis (Duchin, et al., 2010; Chan, et al., 2011; Elyasiani, et al., 2010). However, the investigation of liquidity risk transmission between financial and real estate sectors is missing.

6 Main Contribution Addresses a gap in liquidity literature; Estimates liquidity covariance from three dimensions, liquidity depth, tightness and resilience; Verifies the causal effect of liquidity risks between financial institutions and real estate sector ; Updates the evaluation of liquidity risks of financial institutions and real estate market in China;

7 Data Daily data of 140 public real estate companies and 59 financial institutions (21 Banks and 38 non-bank institutions) between 2000 to 2014; Control variables are selected based on related literature and tested by pairwise correlation coefficient. Control variables: domestic CPI, M2, Bank Reserve, volatility of Shanghai Exchange Index; Shanghai Interbank Offered Rate (Shibor); 5 Year Lending Rate; Chinese Residential Land Price Index (Crlpi); Employment growth ratio of Real Estate, and Industrial Climate Index (Investment in real estate development). Data is from DataStream; National Bureau of Statistics of China, the People s Bank of China, and The Institute of Real Estate Studies (IRES).

8 Liquidity Indicators Three Liquidity Indicators: Liquidity Indicators Equations Liquidity Depth (Amihud): the impact of trading volume on price (Amihud, 2002) Illiq iy = 1 D iy D 1 R idy Volume idy (1) Liquidity Tightness (R-spread): the impact of transaction cost (Sarr & Lybek, 2002) R spread = ask i bid i ask i + bid i 2 (2) Liquidity Resiliency (Roll): the impact of market efficiency (Roll, 1984) Roll = 200 cov i (3)

9 Data Liquidity Depth ( ) Liquidity Tightness ( ) Liquidity Depth of Financial Institutions and Real Estate Firms m1 2005m1 2010m1 2015m1 month Re Depth Liquidity Resilience ( ) Fi Depth Liquidity Resilience of Financial Institutions and Real Estate Firms m1 2005m1 2010m1 2015m1 month Re resilience Fi resilience Fi resilience Fi Depth market_index Liquidity Tightness of Financial Institutions and Real Estate Firms m1 2005m1 2010m1 2015m1 month Re Tightness Variation of Stock Returns Fi Tightness Stock Returns of Chinese Financial Institutions and Real Estate Firms m1 2005m1 2010m1 2015m1 month market_index (mean) rt_re (mean) rt_fi Fi Tightness The risk in liquidity depth (Amihud) of financial institutions is lower than real estate firms. The risks in liquidity tightness (rspread) and liquidity resilience (roll) of financial institutions are higher than real estate firms, and turbulent over time. The stock returns of financial institutions and real estate firms vary along with the variation of Shanghai Stock Exchange index.

10 Empirical Result Real Estate firms to Financial Institutions (Equation 4) Fi depth Fi tightness Fi resilience Re depth *** (12.92) Re tightness (-0.24) Re resilience *** (7.97) M2(YoY) (0.53) (-0.89) (0.31) CPI(YoY) ** ** (2.00) (-2.54) (0.78) Shibor ** (-2.09) (0.28) (-0.00) Market Index ** (-0.56) (2.27) (1.08) Lending Rate *** (1.47) (2.67) (1.27) Bank Reserve (-0.14) (-0.25) (-1.29) crlpi ** ** (-0.48) (-2.49) (-2.23) _cons (-0.36) (0.59) (0.85) N liquidity_fi jt = α liquidity_re it = α N m=1 N m=1 liquidity_re i(t m) + β liquidity_fi j(t m) + β N m=1 N m=1 liq Macro j(t m) + ε jt liq Macro i(t m) + ε it Financial Institutions to Real Estate firms (Equation 5) Re depth Re tightness Re resilience Fi depth *** (10.52) Fi tightness (-1.10) Fi resilience *** (8.09) Re Climate Index (0.51) (-1.82) (-0.16) Re Employ Ratio * ** (1.72) (1.57) (-2.30) Crlpi *** (-1.30) (-5.56) (0.19) _cons *** *** (0.26) (7.50) (3.74) N Newey-West estimator (Newey-West, 1987) is chosen due to correlated error terms, and maximum lags computed from T(Andrews,1991), and confirmed by Order Determination test (Nielsen, 2001). t statistics in parentheses * p < 0.05, ** p < 0.01, *** p < (4) (5)

11 Empirical Result The linear interdependencies among multiple time series are estimated by Vector Auto Regression (VAR) and Vector Error Correlation (VEC) model accordingly due to co-integration among variables(sims, 1980; Johansen, 1995) y it = β 0 + β it i i=1 x i(t m) + δ it i i=1 y i(t m) + v it y (6) x it = δ 0 + β it i i=1 y i(t m) + δ it i i=1 x i(t m) + v it x (7) VAR (Equation 6, 7) VAR (Equation 6, 7) VECM (Equation 6, 7) Fi Re depth Fi Re Fi Re Fi depth(m-1) depth *** * Fi tightness(m-1) tightness *** tightness L._ce1 resilience resilience *** (6.68) (2.01) (6.67) (-1.00) (1.10) (4.89) Fi depth(m-2) Fi tightness(m-2) Fi resilience(m-1) *** (0.96) (-0.30) (1.73) (-0.56) (-6.08) (-0.87) Fi depth(m-3) Fi tightness(m-3) Fi resilience(m-2) ** (0.82) (-0.80) (1.42) (0.14) (-3.12) (-0.37) Re depth(m-1) *** Re tightness(m-1) *** Re resilience(m-1) * (-0.59) (4.12) (-0.61) (6.34) (1.94) (-1.59) Re depth(m-2) Re tightness(m-2) Re resilience(m-2) *** (-1.82) (-0.17) (-0.19) (0.60) (-0.96) (-4.18) Re depth(m-3) * ** Re tightness(m-3) *** _cons (2.11) (2.98) (1.23) (3.98) (-0.13) (0.52) N * p < 0.05, ** p < 0.01, *** p < 0.001; Fi depth: liquidity depth of financial institutions; Fi tightness: liquidity tightness of financial institutions; Fi resilience: liquidity resilience of financial institutions; Re depth: liquidity depth of real estate; Re tightness: liquidity tightness of real estate; Re resilience: liquidity resilience of real estate; L_ce1: long run effect; m: month.

12 Empirical Result Causality test (Fi-Re) A liquidity risk feedback loop is in Equation Excluded chi2 df Prob > chi2 Fi depth Re depth * Re depth Fi depth * Fi tightness Re tightness Re tightness Fi tightness 9.754** Fi resilience Re resilience 10.19*** Re resilience Fi resilience * p < 0.05, ** p < 0.01, *** p < 0.001; liquidity depth between financial institutions and real estate firms. It is positive and significant at 0.1 level. The risk in liquidity tightness of financial institutions is positive significant to real estate sector liquidity risk at 0.05 level The risk in liquidity resilience of real estate firms causes the liquidity risk in financial sector at 0.01 level.

13 Summary A bi-directional liquidity loop exists between financial institutions and real estate sector; Liquidity depth and tightness of the two sectors co-move in the short run, and the liquidity resilience of financial institutions affects real estate firms in the long run; The loop of liquidity risks is stronger between banks and real estate sector; Liquidity risk transmission becomes prominent during episodes of market distress.