Skip to main content
SpringerLink
Log in

Uncovering heterogeneous regional impacts of Chinese monetary policy

  • Published:
Empirical Economics Aims and scope Submit manuscript

Abstract

This paper applies causal machine learning methods to analyze the heterogeneous regional impacts of monetary policy in China. The methods uncover the heterogeneous monetary policy impacts on the provincial figures for real GDP growth, CPI inflation, and loan growth compared to the national averages. The varying effects of expansionary and contractionary monetary policy phases on Chinese provinces are highlighted and explained. Subsequently, applying interpretable machine learning, the empirical results show that the credit channel is the main channel affecting the regional impacts of monetary policy. An imminent conclusion of the uneven provincial responses to the “one-size-fits-all” monetary policy is that different policymakers should coordinate their efforts to search for the optimal fiscal and monetary policy mix.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1

Source: National Bureau of Statistics

Fig. 2
Fig. 3
Fig. 4
Fig. 5

Similar content being viewed by others

Notes

  1. Cortes and Kong (2007), Guo and Masron (2014, 2017) used the official benchmark lending rate or monetary aggregates (M1 or M2) to estimate the regional impacts of Chinese monetary policy.

  2. In the recent decade, the Chinese authorities claimed that the economy was entering a “new normal” with a lower growth rate, and changes in economic structure were expected. Hence, different speeds of changes in economic structure may affect the effectiveness of monetary policy in different regions.

  3. In contrast, conventional econometric approaches, such as those discussed by Carlino and DeFina (1998, 1999), typically adopt a two-step approach. Initially, they estimate the impact of monetary policy on individual regions using a VAR model. Subsequently, they regress the estimated impact on different factors. Although time-varying VAR models can also estimate time-varying heterogeneous impacts, they do not address nonlinearities and interactions among different determinants and heterogeneous impacts simultaneously. This omission can potentially introduce biases into the analysis, highlighting the need for a more comprehensive approach to address these complexities.

  4. The rates of the SLF constitute the upper bound of the corridor, and the lower bound is the interest rate of banks’ excessive deposit reserves paid by the PBoC.

  5. In addition, the PBoC also uses some temporary facilities, including contingent reserve arrangements (CRA), short-term liquidity operations (SLO), the targeted medium-term lending facility (TMLF), and the temporary lending facility (TLF).

  6. A robustness check using conventional econometric methods and the same dataset in this study is obtained, but the results suggest that nearly all the determinants (co-founding variables) are insignificant in explaining the heterogeneous regional impacts of the monetary policy, which is less reasonable. More details of the robustness check can be found in footnote 24.

  7. It is different from permutation importance, which is based on the decrease in the effect on the model performance by removing the specific confounding variable.

  8. Shapley value analysis is a method borrowed from coalitional game theory (Shapley 1953), which is a method for measuring the contributions of single players in a game to the total payout. The Shapley value is the average marginal contribution of a single player across all possible coalitions. In the context of this paper, the average marginal contribution of a single confounding variable across all possible combinations of confounding variables will be estimated. The details of the calculation of Shapley value and SHAP can be found in Lundberg and Lee (2017) and Molnar (2019), Chapter 5.9 and 5.10.

  9. The details of the calculation of partial function can be found in Molnar (2019), Chapter 5.1.

  10. For the two confounding variables case, the interactive PDP is used to show the predicted policy impacts for combinations of different values of two specific confounding variables.

  11. Gap variables are employed as outcome variables in this study due to the research question focusing on the heterogeneous regional impacts of monetary policy in comparison to the national average benchmark. While level variables can also serve as outcome variables, it is worth noting that the findings generally align with those obtained using gap variables (the results can be found in Appendix 5).

  12. However, the drawback of using growth rates is that stability conditions for levels (cointegration) are lost.

  13. The discrepancy between the aggregation of provincial level economic data and national series in China has raised concerns among researchers. Koch-Weser (2013) has pointed out that the data discrepancy is “a complex problem”, as the difference varies across provinces and over time. The problem of overstating economic performance has been improved since the 2000s. Meanwhile, He (2011) argued that the provincial level economic data are reliable for growth regressions.

  14. This is also consistent with the shifting from an export-oriented strategy (mainly the output of secondary industry) to promoting domestic consumption (retail sales and services are mainly included in tertiary industry).

  15. Taking reference to Burriel and Galesi (2018) and Dominguez-Torres and Hierro (2020), extra variables, such as log of GDP per capita, unemployment rate, and non-performance loan (NPL) ratio are considered. However, the results suggest that these variables are insignificant (the variables are not important in the SHAP analysis). See Appendix 6.

  16. More details of the Markov switching model can be found in Hamilton (1989). The same methodology has been applied to China in Zheng et al. (2012).

  17. The updated indicator shows that since the outbreak of the Corona crisis, China had maintained an accommodative monetary stance, although some tightening signs occurred in the second half of 2020 when the Chinese economy recovered and picked up gradually.

  18. In order to create valid confidence intervals and causal inference, 4000 individual trees are used in the causal forest. In addition, the sizes of treated and untreated samples are similar. The shares of treated sample are 43.3% and 56.7% for monetary easing and monetary tightening respectively.

  19. See https://www.ft.com/content/61346c8c-0d09-11e6-b41f-0beb7e589515 and https://www.scmp.com/news/china/economy/article/2104789/liaoning-worst-performer-chinas-northeast-lags-behind-countrys.

  20. For a robustness check, if the share of small banks is replaced by the share of large banks, the importance of the variable (with the opposite sign of impacts) is similar.

  21. The results for variable importance are similar if the share of small firms is replaced by the share of large firm (with the opposite sign of impacts).

  22. Similar results occur when the share of exports is replaced by the share of total trades (sum of exports and imports).

  23. As a robustness check, replacing the share of tertiary industry with the share of the secondary industry (i.e., the negative impacts of a higher share in the secondary industry) shows consistent results.

  24. As a robustness check, the methodology in Cortes and Kong (2007), Guo and Masron (2014, 2017) is applied to the same dataset in this study (See Appendix 7). The results show that nearly all the determinants (co-founding variables) are insignificant in explaining the heterogeneous regional impacts of the monetary policy, which is also different from the results in Cortes and Kong (2007), Guo and Masron (2014, 2017).

References

  • Albuquerque B (2019) One size fits all? Monetary policy and asymmetric household debt cycles in US States. J Money Credit Bank 51(5):1309–1353

    Article  Google Scholar 

  • Athey S, Imbens G (2017) The state of applied econometrics: Causality and policy evaluation. J Econom Perspect 31(2):3–32

    Article  Google Scholar 

  • Athey S, Imbens G (2019) Machine learning methods that economists should know about. Ann Rev Econom 11:685–725

    Article  Google Scholar 

  • Athey S, Wager S (2019) Estimating treatment effects with causal forests: an application. Observat Stud 5:37–51

    Article  Google Scholar 

  • Athey S, Tibshirani J, Wager S (2019) Generalized random forests. Ann Stat 47(2):1148–1178

    Article  Google Scholar 

  • Boeckx J, Dossche M, Peersman G (2017) Effectiveness and transmission of the ECB’s balance sheet policies. Int J Cent Bank 13(1):297–333

    Google Scholar 

  • Burriel P, Galesi A (2018) Uncovering the heterogeneous effects of ECB unconventional monetary policies across euro area countries. Eur Econ Rev 101:210–220

    Article  Google Scholar 

  • Carlino G, DeFina R (1998) The differential regional effects of monetary policy. Rev Econ Stat 80(4):572–587

    Article  Google Scholar 

  • Carlino G, DeFina R (1999) The differential regional effects of monetary policy: evidence from the US States. J Reg Sci 39(2):339–357

    Article  Google Scholar 

  • Chakraborty C and Joseph A (2017) Machine learning at central banks. Bank of England Staff Working Paper 674: 32

  • Chen K, Ren J, Zha T (2018) The nexus of monetary policy and shadow banking in China. Am Econom Rev 108(12):3891–3936

    Article  Google Scholar 

  • Chen K, Higgins P, Waggoner DF and Zha T (2016) China pro-growth monetary policy and its asymmetric transmission. NBER Working Paper, No 22650

  • Clarida R, Galí J, Gertler M (1998) Monetary policy rules in practice: some international evidence. Europ Econom Rev 42(6):1033–1068

    Article  Google Scholar 

  • Clarida R, Galí J, Gertler M (2000) Monetary policy rule and macroeconomic stability: evidence and some theory. Quart J Econ 115(1):147–180

    Article  Google Scholar 

  • Cortes BS, Kong D (2007) Regional effects of Chinese monetary policy. Int J Econom Policy Stud 2:15–28

    Article  Google Scholar 

  • Dominguez-Torres H, Hierro L (2019) The regional effects of monetary policy: a survey of the empirical literature. Journal of Economic Surveys 33(2):604–638

    Article  Google Scholar 

  • Dominguez-Torres H, Hierro L (2020) Are there monetary clusters in the Eurozone-the impact of ECB policy. Journal of Policy Modeling 42:56–76

    Article  Google Scholar 

  • Friedman JH (2001) Greedy function approximation: A gradient boosting machine. Ann Stat 29(5):1189–1232

    Article  Google Scholar 

  • Funke M, Mihaylovski P and Zhu H (2015) Monetary policy transmission in China: A DSGE model with parallel shadow banking and interest rate control. HKIMR Working Papers 12/2015

  • Funke M, Li X and Tsang A (2019) Monetary policy shocks and peer-to-peer lending in China. BOFIT Discussion Papers, 23/2019

  • Funke M, Tsang A (2020) The People’s Bank of China’s response to the coronavirus pandemic–a quantitative assessment. Econ Model 93:465–473

    Article  Google Scholar 

  • Funke M, Tsang A (2021) The direction and intensity of China’s monetary policy: a dynamic factor modelling approach. Econom Record 97(316):100–122

    Article  Google Scholar 

  • Furceri D, Mazzola F, Pizzuto P (2019) Asymmetric effects of monetary policy shocks across US states. Region Sci 98:1861–1891

    Article  Google Scholar 

  • Georgopoulos G (2009) Measuring regional effects of monetary policy in Canada. Appl Econ 41(16):2093–2113

    Article  Google Scholar 

  • Girardin E, Lunven S and Ma G (2017) China’s evolving monetary policy rule: from inflation-accommodating to anti-inflation policy. BIS Working Paper No 641

  • Girotti M (2018) The effects of monetary policy on the composition of bank deposits and on loan supply. Rue de la Banque (Banque de France) 59:18

    Google Scholar 

  • Guo X, Masron TA (2014) Regional effects of monetary policy in China—the role of spillover effects. Asian Acad Manag J 19(1):113–146

    Google Scholar 

  • Guo X, Masron TA (2017) Regional effects of monetary policy in China: evidence from China´s provinces. Bull Econ Res 69(2):178–208

    Article  Google Scholar 

  • Hamilton JD (1989) A new approach to the economic analysis of nonstationary time series and the business cycle. Econometrica 57:357–384

    Article  Google Scholar 

  • Hauptmeier S, Holm-Hadulla F and Nikalexi K (2020) Monetary policy and regional inequality. ECB Working Paper No 2385

  • He Q (2011) Dances with Chinese data: are the reform period Chinese provincial panel data reliable?” Munich Personal RePEc Archive (MPRA) Paper, No 35418

  • Horst M and Neyer U (2019), The impact of quantitative easing on bank loan supply and monetary policy implementation in the Euro area”, Düsseldorf Institute for Competition Economics (DICE) Discussion Paper, No. 325

  • Kamber G and Mohanty MS (2018) Do interest rates play a major role in monetary policy transmission in China?” BIS Working Paper, No. 714 Basel

  • Kerola E and Mojon B (2021) What 31 provinces reveal about growth in China. BIS Working Paper, No 925, Basel

  • Koch-Weser, I.N. (2013), The reliability of China’s economic data: An analysis of national output. U.S.-China Economic and Security Review Commission Staff Research Project

  • Künzel SR, Sekhon JS, Bickel PJ, Yu B (2019) Metalearners for estimating heterogeneous treatment effects using machine learning. PNAS 116(10):4156–4165

    Article  Google Scholar 

  • Lundberg SM, Lee S (2017) A unified approach to interpreting model predictions. Adv Neural Inform Process Syst (NIPS) 17:30

    Google Scholar 

  • Molnar C, Casalicchio G, Bischl B (2020) Interpretable machine learning–A Brief history, state-of-the-art and challenges. In: Koprinska, I et al (Eds.) ECML PKDD 2020 Workshops. ECML PKDD 2020. Communications in Computer and Information Science, Springer: London. p 1323. doi https://doi.org/10.1007/978-3-030-65965-3_28

  • Molnar C (2019), Interpretable machine learning: a guide for making black box models explainable, https://christophm.github.io/interpretable-ml-book/.

  • Mullainathan S, Spiess J (2017) Machine learning: An applied econometric approach. J Econom Perspect 31(2):87–106

    Article  Google Scholar 

  • Pizzuto P (2020) Regional effects of monetary policy in the US: An empirical re-assessment. Econom Lett 190:109062

    Article  Google Scholar 

  • Shapley LS (1953) 17. A value for n-person games. In: Kuhn HW, Tucker AW (eds) Contributions to the theory of games (AM-28), Volume II. Princeton University Press, NJ, pp 307–318. https://doi.org/10.1515/9781400881970-018

    Chapter  Google Scholar 

  • Sun R (2018) A narrative indicator of monetary conditions in China. Int J Cent Bank 14(4):1–42

    Google Scholar 

  • Tiffin A (2019) Machine learning and causality: The impact of financial crises on growth. IMF Working Papers. https://doi.org/10.5089/9781513518305.001

    Article  Google Scholar 

  • Wynne MA, Koech J (2012) One-size-fits-all monetary policy: Europe and the US. Federal Res Bank Dallas Econom Lett 7:9

    Google Scholar 

  • Zheng T, Wang X, Guo H (2012) Estimating forward-looking rules for China’s monetary policy: a regime-switching perspective. China Econ Rev 23:47–59

    Article  Google Scholar 

Download references

Acknowledgement

The author would like to thank Professor Michael Funke (Hamburg University, and Tallinn University of Technology), two anonymous reviewers and the editor for helpful comments on an earlier draft. The usual disclaimer applies.

Author information

Authors and Affiliations

  1. ASEAN+3 Macroeconomic Research Office (AMRO), Singapore, Singapore

    Andrew Tsang

Authors
  1. Andrew Tsang
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Andrew Tsang.

Ethics declarations

Conflict of interest

I confirm that I have no conflicts of interest to disclose for this paper.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Below is the link to the electronic supplementary material.

Supplementary file1 (DOCX 2776 KB)

Rights and permissions

Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Tsang, A. Uncovering heterogeneous regional impacts of Chinese monetary policy. Empir Econ (2024). https://doi.org/10.1007/s00181-024-02575-2

Download citation

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1007/s00181-024-02575-2

Keywords

JEL Classification

Search

Navigation