Until 2018, the US-China trade data gap was in line with the discrepancies found in the bilateral trade data of China and its other partners, but in 2018 the gap began to narrow. This column argues that higher and broader US tariffs on Chinese imports resulted in increased tariff avoidance and an underestimation of US imports from China, which played a significant role in narrowing the US-China trade data gap. Changes in China’s VAT regime may have also played a role.

Tariffs, like all taxes, create distortions. The 2025 US tariff increases are expected to distort global trade flows (Kawasaki 2025) and the pattern of US consumption (Clausing and Obstfeld 2025). Less appreciated is how tariffs can distort US trade data. The 2018-2020 US-China trade war provides a cautionary tale.

From 2009 through 2017, the US recorded receiving imports from China worth 17% more than China reported sending to the US. This bilateral trade data gap was large but steady and predictable. Re-exports from Hong Kong that were correctly recorded as coming from mainland China likely accounted for some of the data gap, but under-reporting of exports at the Chinese border was also likely a factor due to China’s de facto export taxes (Ferrantino et al. 2012). Until 2018, the US-China trade data gap was in line with the discrepancies found in the bilateral trade data of China and its other partners.

There are many reasons why bilateral trade data may not agree, not all of which are due to malfeasance (Levinson and Kellenberg 2016). For example, from 2009 through 2024, China’s reported exports to Germany worth about $48 billion less each year than Germany recording importing from China. During the same period, China’s reported exports to the Netherlands exceeded the value of imports the Netherlands recorded receiving from China by about $44 billion on average each year. Most likely, some Chinese exports to Germany were mis-recorded as going to their first port of call in Europe.

What was unusual was that in 2018 the US-China trade data gap began to narrow. By 2020 the sign on the error term had even reversed. Clark and Wang (2021) highlight two concurrent policy changes that might explain this shift: a decline in China’s net value-added tax (VAT) rates on exports and the increase in US tariff rates on Chinese imports.

Figure 1 US imports from China and Chinese exports to the US

Source: ASR Ltd./LSEG Datastream/national sources

Most countries with value-added taxes fully rebate the VAT on exports. China’s VAT rebate rates have historically been set below its VAT rates, meaning it effectively taxes exports. From 2017 to 2020, China lowered its gross VAT rates and increased its VAT rebates. Its net VAT rates on exports fell from about 7% to 2.5%. These policy changes were partly designed to offset the impact of the concurrent US section 301 tariff hikes, which raised the effective tariff rate on China from about 3% to an average rate of about 19%.

Because China’s VAT changes affected its exports to all trade partners while the tariffs only impacted US imports, Clark and Wong were able to distinguish between the impact on the US-China trade data gap from each policy change. The trade data gap swung by $88 billion from 2017 through 2020 in absolute terms. They found that $55 billion of the swing was due to US tariff avoidance and $12 billion was due to overreporting (or a decrease in underreporting) of Chinese exports.

My product-level analysis of the US-China bilateral trade data also suggests tariff avoidance played a significant role in narrowing the trade data gap (Wolfe 2024). I mapped China’s exports and US imports by six-digit HS code to the US Trade Representative’s Section 301 tariff list. The exports and imports were then sorted into two buckets: those subject to the Section 301 tariffs and those not.

Both the Chinese export data and the US import data show the expected decline in shipments of products subject to the tariffs between 2017 and 2020, but the decline in the US import data is much steeper ($75 billion according to the US, $10 billion according to China). China reported a modest increase in its exports of non-tariffed goods over this period, whereas the US reported its imports of those products were virtually flat.

The sign on the trade gap only changed for the products subject to Section 301 tariffs in 2020, suggesting tariff avoidance accounted for most of the shift. The smaller narrowing of the gap for other products suggest changes in China’s VAT regime may have also played a role. Interestingly, the sign on trade gap for these non-tariffed products also changed in 2023.

Figure 2 US imports and Chinese exports by US Section 301 tariff regime

Source: ASR Ltd./Comtrade/US Trade Representative

The shift in China’s trade data gap with the rest of the world mirrored the change in its trade data gap with the US for the basket of goods not subject to the Section 301 tariffs. Figure 3 compares China’s trade data gap with its next 15 largest export markets  and the overall gap with the US. The data gap between China’s exports and its partners’ imports narrowed between 2017 and 2020. China’s reported exports were worth 8% less than its partners recorded receiving in 2017 but only 0.5% less in 2020. This shift may have been due to the changes in China’s VAT regime, although similar shifts in the trade data gap have occurred before. Interestingly, the sign on the trade data gap also reversed from 2021 through 2023, albeit at a considerably lower level than seen in the bilateral data with the US.

Figure 3 China’s bilateral trade gap

Source: ASR Ltd./LSEG Datastream/national sources

In 2025, the average US tariff rate on Chinese goods increased again, reaching 57.6% in September (Bown 2025). This further increase in tariff rates coincided with a further increase in the US-China bilateral trade gap. In the three months through July, China reported exporting goods worth 55% more than the US reported importing from China. There has not been a similar shift in China’s trade data gap with its next 15 largest export markets. This suggests that higher and broader US tariffs on Chinese imports resulted in increased tariff avoidance and an underestimation of US imports from China.

The official statistics cannot show how companies have managed to avoid paying US tariffs. But the scope for re-routing, transfer price adjustments, country of origin re-labelling, and so on appears to be substantial. After all, nearly 50% of US imports happen within the same company according to Customs data (Ruhl 2015).

The implications are twofold. First, this analysis suggests that US import data overstate the degree of de-coupling from China since 2018. In this sense, tariffs may not distort trade flows as much as the official data suggest. Second, it seems likely that the measurement error in US import data is no longer confined to China. In 2025, the average US tariff rate on the world ex-China was raised to 19.5%, roughly equal to the level imposed on China during the 2018-2020 trade war (Bown 2025). Further research may find that US trade data gap with its other major trade partners will shift accordingly if the tariffs remain in place.

Source : VOXeu