Official vs. effective, real vs. nominal, bilateral vs. multilateral.
First, let’s the nominal as compared to the inflation adjusted.
Figure 1: Official bilateral nominal USD/CNY exchange rate, period average (blue, left scale), log bilateral real exchange rate (tan). NBER defined peak-to-trough recession dates shaded gray. “Down” is a CNY depreciation. Source: FRED, NBER, and author’s calculations.
Second, let’s consider the difference between the official exchange rate, and the rate that transactions are conducted at. In advanced economies, they are typically the same; not so for many developing countries. In China this distinction becomes important insofar as the observed “mega-devaluation” in 1994 was in some ways an artifact of using the official exchange rate. Fernald, Edison and Loungani (1999) note that many transactions were conducted through swap centers. Using their series yields a less dramatic drop in 1994.
Figure 2: Official bilateral real USD/CNY exchange rate, period average (tan), and bilateral exchange rate adjusted for swap center rates used before 1994 (red). NBER defined peak-to-trough recession dates shaded gray. “Down” is a CNY depreciation. Source: FRED, Fernald, Edison, Loungani (JIMF, 1999), NBER, and author’s calculations.
Two aspects are of interest here. First, the trend CNY appreciation using the official rate is essentially zero. However, using the swap center adjusted rate, the CNY has been appreciating about 1% per annum over this period (estimated using first differences).
Second, the real exchange rate looks possibly trend stationary if one uses the official rate (rejects Elliott-Rothenberg-Stock unit root test at 10%, but also rejects Kwiatkowski-Phillips-Schmidt-Shin trend stationary test at 1%), but is pretty clearly I(1) using the adjusted rate (fail to reject ERS, and rejects KPSS).
Third, and particularly relevant now (see Sobel, 2022), the bilateral and real exchange rates can differ substantially in terms of direction.
Figure 3: Bilateral exchange rate adjusted for swap center rates used before 1994 (red), and trade weighted value of CNY (teal), both in logs, 2000M01=0. Trade weighted CNY is against broad basket of currencies. NBER defined peak-to-trough recession dates shaded gray. “Down” is a CNY depreciation. Source: FRED, Fernald, Edison, Loungani (JIMF, 1999), BIS, NBER, and author’s calculations.
What becomes clear is that the recent depreciation of the CNY is really a depreciation against the dollar. Once one takes a trade-weighted view, the CNY is about at the same (inflation-adjusted) level it was in 2015.
Addendum, 9/9/22, 8AM:
As for whether the CNY is misaligned, there are numerous ways of answering this question. The most appropriate depends on the economic model used, the data being used, and the statistical characteristics of that data. For more, see Cheung, Chinn, and Fujii (JIMF, 2007). Using PPP and monetary models, Chinn (2000) evaluated exchange rates other than the Chinese on the eve of the 1997 currency crises.