Price-based Monetary Policy and the Estimation of the Natural Interest Rate in China Li Hongjin Su Naifang Research Division, Operations Office, the People s Bank of China BIS-CAFRAL Ninth Annual Workshop of the Asian Research Networks, March 2016,Mumbai The views of this paper are those of the authors and do not necessarily represent the PBC
Outline Why to estimate the Natural Interest Rate(NIR) in China r* as the benchmark for the success of monetary policy Monetary policy from quantity- to price- framework How to estimate NIR Semi-Structural space state model of Laubach & Williams(2003) What the NIR looks like The typical facts of the return on capital The role in economic and monetary analysis The determinants, i.e. tech., labor, preference
Ⅰ. Background Deposit rate ceilings canceled and the nearly 20-year interest liberalization reform almost fulfilled in Oct. 2015 Still manage the base rate of loan & deposit due to the weak banks pricing ability a market orientated benchmark interest operation and its transmission mechanism challenges to the traditional quantity measures and increasing effectiveness of price policy with the interest rate liberalization NIR as the benchmark for the success of monetary policy Major central banks experience: Returned to interest rate dominance monetary framework in mid-1980s, Taylor s Rule, r* Ferguson(2004) Disputes between Bernanke(2010, 2015) &Taylor(2010,2015) FED liftoff and normalization, Yellen(2015a,b)
Ⅱ. Monetary Policy Implementation in China: From quantity-based framework to price-based framework Monetary policy with direct controls Monetary policy originated from the planed economy Credit quota controls and cash issuing as the intermediate and operational targets Central bank lending and rediscounting as the main policy instruments and base money supply channel Turned to indirect quantity framework in the mid-1990s M2 as the intermediate target together with RMB credit & loan increment at some time Excess reserves and money market rate as operational targets OMO as the main policy instruments and base money supply channel focusing on base money and market liquidity Together with RRR, rediscounting, central bank lending and central bank s base rate Only about 20-year history of indirect monetary policy in modern sense
Excess liquidity abundance period from 2002 to 2012 To issue the central bank bills to sterilize the liquidity Prone to quantity measures with 41 times of RRR adjustments and 22 times of base rates adjustments
Tuning to the price-based monetary policy Liquidity environment changed due to BOP & exchange rate approaching to equilibrium after 2011, and the FX reserves and capital began to decrease and outflow Rapid growth of financial innovation & disintermediation with interest rate liberalization More and more sophisticated financial system Unstable money multiplier and velocity
Strengthening the price-based monetary policy Decreasing effectiveness of quantity policy and money market fluctuations in the mid-2013 Innovative liquidity instruments to enhance liquidity and interest rate management 2013, SLO, OMO every day, Feb., 2016 2013, SLF, as the interest rate corridor in Nov. 2015 2014, MLF as the mid-term liquidity management 2014, PSL, a supplement longer-term credit instrument and modified central bank lending
Innovative Liquidity Instruments (Billion RMB)
Main market /central banking rates
Ⅲ. Model of NRI Estimation, Data & the Results Estimation methods: DSGE v.s. LW(2003) DSGE: structural full-blown model with strong theoretical restrictions, but too complicated, more prone to misspecification and sensitive to parameters DSGE: short-run view of Wicksellian NIR with more fluctuations (Kiley, 2015) LW state space model with New Keynesian: semi-structural model, still with theoretical restrictions and explore the information on data LW: longer-run view, equilibrium real interest rate (Kiley, 2015) The two methods are complementary to each other and are used at the same time when making policy (LW, 2015; Yellen, 2015b)
Model Structure: Measurement Equations: IS Curve: Phillips Curve: Monetary Reaction Function/ Taylor s Rule: Where: output gap Transition Equations: Ramsey Euler equation: Assuming:, ex ante real rate,, CPI:
Measurement Equations: Transition Equations: Trend analysis and OLS to achieve the initial parameters, and then use EM algorithm with Kalman filter method and MLE to get the unobserved variables and the parameters
Data 2004Q1 2015Q2 quarterly data 2010 price constant logarithm GDP seasonally adjusted Overnight repo rate Labor force of the population ages of 15-64 Ex ante real rate by VAR of GDP, nominal interest rate, CPI, labor growth
Estimated NIRs with and without monetary target and economic growth 2004-2006:3.54%(±0.65%);2007-2011: 2.34%(±0.70%);2012-: 1.93%(±0.26%) More smooth with a monetary target with the same average(irrelevant of monetary policy framework, real economy concept) Granger Causality of GDP Table1: VAR Granger Causality/Block Exogeneity Wald Tests of NIR and GDP Growth Dependent variable: NR Dependent variable: GDP Excluded Chi-sq df Prob. Excluded Chi-sq df Prob. GDP 5.45 4 0.6534 NR 36.61 4 0.00
Ⅳ. Typical Facts of the Return on Capital and Further Analysis of NIR in China NIR and Return on capital in China Myrdal(1939) Lindahl(1939) consistent with the marginal productivity of capital, i.e. the return on capital; the concept derived from the real economic activity saving equals to investment; determined by the supply and demand of capital stabilizes the price level; the final effect of NIR Woodford(2003); Amato( 2005)
To achieve one-year NIR by expectation theory Bai, et al.(2006), Bai and Zhang(2014) have provided the yearly data of return on capital from 1978-2013 Table2: Equality Test between NIR and Return on Capital Quarterly Data Yearly Data Method Statistics P-value Statistics P-value Equality of Means t-test -0.87 0.38-0.55 0.59 Equality of Medians Wilcoxon/Mann-Whitney 0.60 0.55 0.04 0.97
Rate Gap in Economic Analysis Estimate the output gap jointly Significantly and negatively correlated with output gap and CPI (countercyclical) and Granger causality with each other, an indicator to judge the economic situation Table3: VAR Granger Causality/Block Exogeneity Wald Tests of Rate Gap, Output Gap and CPI Dependent variable: CPI Dependent variable: Rate Gap Dependent variable: Output Gap Exclu ded Rate Gap Outpu t Gap df Prob. Excluded df Prob. Exclu ded Chisq Chisq Chisq 29.12 1 0.00 CPI 8.73 1 0.00 CPI 7.94 1 0.00 26.79 1 0.00 Output Gap 5.71 1 0.02 Rate Gap df Prob. 374.2 1 0.00 All 29.59 2 0.00 All 11.58 2 0.00 All 432.1 2 0.00
Rate gap and monetary policy adjustments significantly and negatively correlated with RRR adjustments Gragner causality of RRR adjustments, as the anchor of interest rate decision Evaluation of monetary policy stance: in theory, if r>r*, restrictive; if r<r*, accommodative; However, due to interest control and financial depression, r<r* in most cases, then when the rate gap 0, restrictive; else,accommodative Table4: VAR Granger Causality/Block Exogeneity Wald Tests of Rate Gap and RRR Adj. Dependent variable: Rate Gap Exclude Chisq. Prob df d RRR 3.56 2 0.17 Adj. Dependent variable: RRR Adj. Exclude Chisq. Prob df d Rate 14.39 2 0.00 Gap
Determinants of NIR Ramsey Euler equation: Significantly correlated with tech. growth and labor growth Preference unchanged
V. Conclusion Remarks Estimate China s NIR based on LW(2003) with a monetary reaction function including monetary target and labor factor in Euler equation Average NIR is 2.5%, consistent with the return on capital and closely/granger causally correlated with economic growth Though smoother than the outcome without monetary target, NIR is irrelevant to monetary framework Rate gap is countercyclical: an indicator to judge the economic situation and can be used as the benchmark of monetary policy decision and evaluation Preference unchanged. Due to the lowered but more stable tech. growth and decrease of labor growth, NIR will decrease moderately in the future The price-orientated monetary framework in the near future: clear a specific target policy rate, guide the rate level through OMO and interest rate corridor arrangements based on NIR and output and inflation gap to fulfill a stable economic growth and inflation
Thank you and comments welcome!