The data used in this paper are sourced from the ASEANstats database. The sample covers all ten ASEAN member states (Brunei Darussalam, Cambodia, Indonesia, Laos, Malaysia, Myanmar, the Philippines, Singapore, Thailand and Viet Nam). Bilateral flows are measured as ASEAN’s imports from China and market size as ASEAN’s imports from the world, both recorded on the ASEAN side in current US dollars at the two-digit level of the HS 2017 nomenclature (chapters 01 - 24); because the decomposition operates on chapter-level aggregates, which are stable across HS revisions, no recoding was required, and missing observations were treated as zero. In the CMS calculations ASEAN is treated as a single aggregated market, i.e., the summation over the market dimension j collapses to the ten-member total, and country-level shares are used only descriptively in Table 1. Each sub-period’s decomposition is constructed by the annual method, summing year-on-year component changes within the sub-period. Taking the launch of the RCEP negotiations in 2012, the signing of the Upgrade Protocol of the China-ASEAN Free Trade Area in 2015, and the formal entry into force of RCEP in 2022 as key time nodes, the period 2006 - 2025 is divided into four sub-periods: 2006 - 2012, 2013 - 2015, 2016 - 2020, and 2021 - 2025. The scope of agricultural products follows the first twenty-four chapters of the HS nomenclature, covering four categories: animal products (HS 01 - 05), plant products (HS 06 - 14), animal and vegetable oils and fats (HS 15), and processed agricultural products (HS 16 - 24).

The Constant Market Share Model (CMS) was first proposed by [8]. The model assumes that if a country’s export competitiveness in a given commodity category remains unchanged, its export share in the international market stays constant; on this basis, it decomposes a country’s export growth into changes in the structure of market demand and changes in competitiveness, namely the structural effect, the competitiveness effect, and the second-order effect. [4] further decomposed export growth into the growth effect, product effect, overall competitiveness effect, specific competitiveness effect, pure second-order effect, and dynamic second-order effect, forming a complete two-layer decomposition framework. The first-layer decomposition is specified as follows:

The second-layer decomposition further disaggregates export growth across the product and market dimensions:

where the last term on the right-hand side of the second-layer decomposition is written as:

The simplified CMS model (third layer) is then obtained as:

In the above expressions, q denotes China’s agricultural export value to ASEAN; Q denotes ASEAN’s total agricultural import value; S denotes the share of China’s agricultural exports in ASEAN’s total agricultural imports; subscripts i and j refer to the product category and the importing market respectively, so that the double subscript ij denotes a product-market pair; superscripts 0 and 1 denote the base period and the end period; Δ denotes the change of a variable from the base period to the end period; the summation over i runs across all product categories, and the double summation over i and j across all product-market pairs; and Ξ collects the second-order (interaction) terms appearing on the right-hand side of the second-layer decomposition. In the first-layer decomposition, the structural effect reflects changes in ASEAN’s market demand, the competitiveness effect reflects changes in the competitiveness of China’s agricultural products, and the second-order effect reflects the interaction between the two. This paper adopts the annual aggregation method in computing the decomposition to eliminate the influence of intra-period share fluctuations.

In the simplified CMS model, the structural effect is split into the growth effect and the product effect, the competitiveness effect into the overall competitiveness effect and the specific competitiveness effect, and the second-order effect into the pure second-order effect and the dynamic second-order effect.

The economic meanings of the effects are summarized in Table 2.

Table 2.The CMS decomposition and corresponding interpretations.

The first-layer decomposition of the CMS model is reported in Table 3.

Table 3.First-layer CMS decomposition of China’s agricultural exports to ASEAN (100 million USD).

Note: figures in parentheses denote the share of each effect in the actual change in exports.

The structural effect is a key driver of the fluctuation in China’s agricultural exports to ASEAN. In 2006 - 2012, its contribution reached 101.31%, far exceeding the competitiveness effect and the second-order effect, indicating that export growth in this stage was mainly driven by the expansion of ASEAN market demand. After the launch of the RCEP negotiations in 2013 - 2015, the contribution of the structural effect fell sharply to 42.47%, while the competitiveness effect rose to 57.05% and became the leading factor. In 2016 - 2020, with the gradual implementation of the Upgrade Protocol of the free trade area, the structural effect recovered to 60.32%, though it no longer exerted the overwhelming influence as before. In 2021 - 2025, after RCEP formally entered into force, the contribution of the structural effect rose further to 70.53%, indicating that import demand in the ASEAN market remains the principal driving force behind the growth of China’s agricultural exports.

The competitiveness effect exhibited a pattern of “sharp rise followed by gradual decline” over the sample period: its contribution jumped from 6.04% in 2006 - 2012 to 57.05% in 2013 - 2015, then fell back to 38.02% in 2016 - 2020 and 30.55% in 2021 - 2025. In absolute terms, however, it increased from 2.70 to 4.64, 13.98, and 23.65 (100 million USD) in successive periods, equivalent to annual averages of 0.39, 1.55, 2.80 and 4.73 (100 million USD) over the 7-, 3-, 5- and 5-year sub-periods, indicating that the competitiveness of China’s agricultural products in the ASEAN market has continued to strengthen in magnitude, and that the declining contribution mainly stems from the expanding base of the actual change in exports in each period. The contribution of the second-order effect stays within 2% in all periods except 2006 - 2012 (−7.35%); its influence is limited and is not discussed further.

The second-layer decomposition of the CMS model is reported in Table 4.

Table 4.Second-layer CMS decomposition of China’s agricultural exports to ASEAN (100 million USD).

Note: figures in parentheses denote the share of each effect in the actual change in exports. With ASEAN treated as a single aggregated market and the annual method, the dynamic second-order term vanishes identically, so the second-order effect reported here is entirely the pure second-order effect defined in the model.

Within the structural effect, the growth effect dominates. Its contribution reached 96.03% in 2006 - 2012, indicating that before the launch of the RCEP negotiations China’s agricultural exports to ASEAN were highly dependent on changes in the importing side’s demand. In 2013 - 2015, the growth effect dropped sharply to 44.39%, showing more autonomous competitiveness in China’s agricultural exports. It recovered to 62.30% in 2016 - 2020, a period during which the Upgrade Protocol entered gradually into force, although the decomposition itself cannot identify a causal effect of the Protocol. In 2021 - 2025, after RCEP took formal effect, the contribution of the growth effect rose further to 77.94%, indicating that the expansion of ASEAN import demand remains the primary driver of the growth of China’s agricultural exports. The contribution of the product effect has been consistently low or even negative, indicating a poor match between China’s export structure and ASEAN’s import structure.

Within the competitiveness effect, the overall competitiveness effect contributes little, while the specific competitiveness effect plays the leading role. In absolute terms, the specific competitiveness effect grew from 2.60 (100 million USD) in 2006 - 2012 to 4.26 in 2013 - 2015, 12.86 in 2016 - 2020, and 22.14 in 2021 - 2025; its contribution jumped from 5.82% to 52.42% in 2013 - 2015 and then fell back to 34.97% and 28.60%. The surge in 2013-2015 coincided with the launch of the RCEP negotiations; whether it reflected geopolitical and price advantages in that period is a contextual interpretation rather than a finding of the decomposition; the subsequent decline in the contribution again reflects the expanding export base rather than any absolute deterioration in competitiveness.

Agricultural products are divided into four categories by HS code—animal products, plant products, animal and vegetable oils and fats, and processed agricultural products—for which separate CMS models are estimated; the results are reported in Table 5.

Table 5.CMS decomposition of China’s agricultural exports to ASEAN by product category (100 million USD).

Note: figures in parentheses denote the share of each effect in the actual change in exports of the corresponding category.

In terms of the growth effect, its contributions for animal products and plant products were generally higher than those for animal and vegetable oils and fats and processed agricultural products. A plausible interpretation is that ASEAN, a region of numerous archipelagos at low latitudes, cannot be self-sufficient in animal and plant products, whereas China is geographically close to ASEAN and richly endowed with agricultural product varieties, giving these two categories a marked comparative advantage in the ASEAN market. For animal products, the growth effect fell from 75.12% in the first stage to 29.96% in the second, rebounded sharply to 288.55% in the third, and stood at 73.10% in the fourth; for plant products it fluctuated from 112.22% to 115.58% in the fourth stage.

In terms of the product effect, the contributions are low or negative for three of the four categories, indicating a poor match between the structure of China’s agricultural exports and ASEAN’s import structure. The product effects of animal products and processed agricultural products are negative in most periods, revealing a mismatch between the export structure and import demand; plant products are the exception: their product effect has been positive in every sub-period since 2013 - 2015 (4.03, 2.82 and 3.26, in units of 100 million USD), although its share in the actual change in exports fell from 109.33% to 12.89% as the export base expanded.

In terms of the competitiveness effect, the effects for animal products and plant products fluctuated the most. For animal products, the effect was positive in 2006-2012 and 2013 - 2015 (with contributions of 22.24% and 90.01%, respectively), turned negative (−182.85%) in 2016 - 2020, a period of pronounced political and international disturbances, though the decomposition cannot establish these factors as causes, and recovered to a positive value (26.70%) in 2021 - 2025. For plant products, the effect was negative in 2006 - 2015 (−8.10% and −80.64%), turned positive and rose to 38.47% in 2016 - 2020, and fell back to −28.46% in 2021 - 2025, with fluctuations clearly larger than those of the other categories. The competitiveness effects of animal and vegetable oils and fats and of processed agricultural products strengthened markedly during 2021 - 2025, reaching 56.21% and 47.37%, respectively, indicating sustained gains in China’s competitiveness in these two categories.