Manipulating export and import transaction data for capital control evasion is an old trick but difficult to detect. This paper provides new empirical evidence by taking advantage of the unique institutional setting of RMB’s dual exchange rates in mainland China and Hong Kong. We find that the spread of the RMB-USD exchange rates in mainland China and Hong Kong is an important driving force for monthly fluctuations in the aggregate trade data gap, which is defined as discrepancies between the bilateral trade data reported by mainland China and Hong Kong respectively. This finding indicates capital flows camouflaged under the trade account for exchange rate arbitrages. We further adopt the Benford’s law test (BLT) to disaggregate trade data of mainland China and Hong Kong to identify the products prone to data manipulations. For the group with a higher share of the products that fail the BLT, we find a similar pattern between the exchange rate spread and the aggregate trade data gap, but not for the group of products that pass the BLT. Our results highlight the challenges to implementing capital controls and can help improve the effectiveness of the policy and reduce its associated costs.