Calibration of autoregressive parameter of U.S.' and China's inflation objective shock

Estimation ML/Bayesian estimation
1

Dear Johannes and all,
I am formulating and estimating a two-country DSGE model for U.S. and China, and in each country, I model inflation objective as an autoregressive shock. However, I am not sure about how to calibrate autoregressive parameters for U.S.’ and China’s inflation objective shock, should I set autoregressive parameter of inflation shock to 0.9 for both U.S. and China? Is it 0.9 too small? Should U.S.’ calibrated autoregressive parameter of inflation shock be smaller than that of China or the rother way around?
Thank you very much and look forward to hearing from you.
Best regards,
Jesse

2

If it is like a course project or a thesis then I think you can put whatever value for Chinese inflation objective.
… But not in a paper for publication. Simply because China doesn’t have an inflation objective. The goal of Chinese monetary policy is “to keep the economy performing within the appropriate range”, which means they are not looking for an autoregressive goal at all.

3

Thank you for your reply, but for China, the inflation target is changing every year, that is why I think time-varying inflation target may be applicable in China. According to the website (How's China's inflation measured and why is it important? | South China Morning Post), it says “China has set a CPI growth target of around 3 per cent per cent, compared to around 3.5 per cent last year. The projected targets for growth, employment, and CPI should keep the economy performing within the appropriate range.”, that is why I think using a time-varying inflation target is better than a constant inflation target in China’s monetary policy rule.

4

This working paper 'BIS Working Papers No 641 China’s evolving monetary
policy rule: from inflation-accommodating to anti-inflation policy
work641.pdf (475.6 KB)
’ also indicates that China monetary policy rule looks like to be characterized by informal inflation targeting.

5

Well, without explicit monetary policy rules that means the rules are potentially arbitrary, while in practice I believe the PBoC (Chinese central bank) carries out some non-independent inflation targeting policies.
In practice, you could just estimate an autoregressive process coming from data, but if you ask about explicit policy shocks then my answer is – you can claim them, though not really.

1 Like
6

Let me add a couple of thoughts:

  1. Central bankers are obviously more complex and simple monetary policy rules. But we employ such rules in our models because they have been shown to often quite well describe actual central bank behavior. Thus, it’s irrelevant if there is a formal inflation target if the central bank nevertheless behaves as if.
  2. You are trying to describe actual central bank behavior. If you don’t know how to do that, picking parameters is not a good idea. You may indeed want to estimate the process.
  3. That being said, an autoregressive parameter implies that there is a long-run inflation target. Sometimes, it’s more appropriate to model the inflation target as a unit root process.