Foreign exchange intervention and monetary policy rules under a managed floating regime: evidence from China

ABSTRACT

This paper investigates the consequences of foreign exchange (FX) intervention in monetary policy under a managed floating regime. Focusing on China’s FX intervention, we identify periods of strong and weak FX intervention using a Markov regime switching approach. We then evaluate quantity-based monetary policy rules using both regime switching reduced-form and structural estimations. In particular, monetary policy regimes obtained from the structural estimation match well with the previously identified intervention regimes. We find that the People’s Bank of China has significant exchange rate stabilization incentives during periods of strong FX intervention, and that the monetary policy rule depends on the state of FX intervention. Furthermore, our estimations point to a trade-off between the central bank’s internal and external policy targets in that strong FX intervention leads to weak responses to domestic GDP and inflation fluctuations.

KEYWORDS:

• Managed floating
• FX intervention
• monetary policy
• China

JEL CLASSIFICATION:

• E52
• E58
• F31
• F33

Acknowledgments

We would like to thank Gianluca Benigno, Kaiji Chen, Yiping Huang, Valerio Nispi Landi, Tao Liu, Xingwang Qian, Barbara Rossi, Jay Shambaugh, Changhua Yu, Vivian Yue and Bo Zhao for very helpful comments. Many useful remarks were also provided to the authors during the seminars at Indiana University, Simon Fraser University, Peking University, as well as at the 5th HenU/INFER Workshop on Applied Macroeconomics, 2019 China Economist Society (CES) Conference, 2019 Asian Meeting of the Econometric Society, 50th Anniversary Conference of the MMF Research Group at London School of Economics, 2019 Singapore Economic Review Conference and 2019 Midwest Macro Meeting.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 Ilzetzki, Reinhart, and Rogoff (2019) document an increasing share of countries with limited exchange rate flexibility and a record accumulation of reserves after 2002. For recent discussions on the trilemma, see Obstfeld, Shambaugh, and Taylor (2005), Aizenman, Chinn, and Ito (2010) and the references therein.

2 See the PBC’s quarterly Monetary Policy Report (MPR) in 2005Q4, 2006Q2, 2007Q2, 2013Q3, and 2016Q2. MPR is the PBC’s official release that describes how its monetary policy and exchange rate policy respond to economic and financial conditions in each quarter.

3 See the PBC’s MPR in 2011Q4, 2015Q4, and 2017Q3.

4 For instance, Financial Times reported the PBC has spent nearly half a trillion dollars in foreign reserves to intervene in the FX market since August 2015 to prevent RMB from depreciation. See the link https://www.ft.com/content/26484358-308d-11e6-bda0-04585c31b153.

5 It is widely accepted that, in the recent decades, China has adopted a quantity-based monetary policy rule. Chen, Ren, and Zha (2018) estimate the M2-based monetary policy rule with a focus on the pro-growth aspect of Chinese monetary policy using a regime switching method. The exchange rate stabilization is among the central bank’s objectives which is clearly stated in the PBC’s mandates. For recent studies on the RMB exchange rate policy, see Jermann et al. (2019) and Das (2019).

6 Evidence of FX purchases (sales) that increased (decreased) the funds outstanding for foreign exchange and the M2 growth rate can be found in the PBC’s MPR in 2013Q1 (2016Q2).

7 In early 2009, the PBC’s governor Zhou Xiaochuan said in public that ``the choice of RMB’s repeg to US dollar is a response to the global financial crisis which generated large increase in the uncertainty”. See the link: https://www.caixinglobal.com/2016-02-15/transcript-zhou-xiaochuan-interview-101011865.html.

8 Since the year of 2002, the PBC started to allow the private sectors and financial institutions to hold foreign currency that arises from transactions in the current account. In other words, they are not forced to sell these foreign currency holdings to the central bank and are able to hold them as much as they are willing to. See Holden et al. (2021) for more institutional details on China’s interbank FX market.

9 Similar measures of FX intervention strength have been used by Levy-Yeyati, Sturzenegger, and Alfredo Gluzmann (2013) and Frankel (2019).

10 For the plot of the two measures, see in the Appendix.

11 Table A1 in the Appendix summarized the coefficients of correlation and the significance levels between these two alternative measures and funds outstanding for foreign exchange..

12 The PBC turned to indirect FX intervention such as counter-cyclical factors and other macro-prudential tools in the second half of 2017 to keep RMB exchange rate stable. See the PBC’s MPR in 2017Q4.

13 See the detailed description of PBC’s quantity-based monetary policy in Chen, Ren, and Zha (2018), which argues that a quantity-based rule better fits for Chinese data than a price-based rule.

14 A better way to capture the FX intervention is to use higher frequency (daily) data; however, given the data limitations, monthly data are the best that we can obtain for China. We use monthly data to capture as much information as possible.

15 We do not covert the dependent variable into a dollar value here, since doing so would affect the inherent non-linearity of the data. For example, an episode may be mistakenly identified as an intervention because the exchange rate changes, even though the central bank does not take any intervention measures.

16 This differs from the classic structural break test of Chow (1960), which is based on known dates.

17 This is consistent with the findings of Das (2019), who claims that “market sentiment turned in 2014 toward a rising sense that the RMB was becoming overvalued and intervention turned from FX purchases to growing sales.”

18 It would be more intuitive if the new regime started from mid-2015, when the PBC initiated the RMB central parity reform and the large RMB depreciation afterward. However, since structural break tests and regressions are more likely to generate persistent regimes, our classification identified weak selling periods in 2014 as well. The regime switching model we propose in the next section solves this issue.

19 To reduce model complexity, we assume regime 1 does not move to regime 3 directly, and vice versa.

20 We also estimate the monetary policy rule with GDP growth switching from 2005:Q3 to 2017:Q2, and the results are discussed in Section VI.

21 We also test the price-based monetary policy rule using the 7-day Repo instead of M2, see discussions in the Appendix C.:

22 The model is estimated using Junior Maih’s RISE toolbox, an object-oriented MATLAB toolbox for solving and estimating Markov switching rational expectations models, developed by Maih (2015).

23 In policy circles, the debate post-2015 has been about the extent to which China cared about the currency basket versus the USD rate. We also estimate the reduced and structural form of monetary policy using the calculated basket index (based on Frankel and Xie 2010). However, our results show that the RMB/USD exchange rate better fits the monetary policy and FX intervention.

24 If we use the previous method directly on the 2005–2017 sample, we may need more than three states in Markov switching model due to the financial crisis as well as multiple structural changes in exchange rate policy. However, more states in the model mean that there are additional constraints on transition probabilities as well as a loss of degree of freedoms.

25 We also attempt an independent switching on the GDP coefficient, however, its regime switching results are not statistically significant.

²⁶ See https://www.swift.com/insights/press-releases/more-than-100-countries-are-now-using-the-rmb-for-payments-with-china-and-hong-kong for details.

²⁷ It is also important for financial stability in the Asia-Pacific region because the co-movement between the RMB and other currencies depends on the extent of RMB management (Marconi 2018).

²⁸ See PBC’s announcement on July 21, 2005, ``Reforming the RMB Exchange Rate Regime'.' The RMB was allowed to fluctuate with a much wider band against the other four major currencies: EUR, Hong Kong dollar, JPY, and GBP, which showed the importance of the USD for the RMB exchange rate regime.

²⁹ See PBC’s announcement on June 19, 2010, “Further reform the RMB Exchange Rate Regime and Enhance the RMB Exchange Rate Flexibility.”

³⁰ The capital and financial account switched from capital inflows of 346.1 billion USD in 2013 to capital outflows of 51.4 and 485.3 billion USD in 2014 and 2015, respectively. The surge of the capital outflow in 2015 was largely due to short-term investment flows.

³¹ See Shu, He, and Cheng (2015) for a detailed study on the onshore and offshore RMB markets.

Additional information

Funding

The work was supported by the National Natural Science Foundation of China [71903191, 71850009]; National Social Science Foundation of China [19AJY028, 20ZDA053] Fund for building world-class universities (disciplines) of Renmin University of China [KYGJC2021002]; Fundamental Research Funds for the Central Universities [63202064]