Does Indonesian Sustainable Palm Oil Certification Improve Oil Palm Production? Evidence from District-Level Data in Indonesia

2.1 Sustainable palm oil certification and oil palm production

Sustainability certification has increasingly been positioned as an instrument for improving production practices, strengthening market credibility, and promoting more responsible palm oil governance. In the Indonesian context, empirical studies examining the production effects of ISPO certification remain relatively limited. Nevertheless, existing evidence suggests that ISPO may contribute to production performance when certification requirements are translated into improved plantation management and good agricultural practices.

Ernah et al. [11] provided one of the earliest empirical analyses of ISPO-related practices in Indonesia. Using household-level data from nine districts across provinces in Sumatra and applying a random-effects regression model, they found that the number of ISPO practices adopted by smallholder households was positively associated with oil palm production per hectare. Rahutomo et al. [12] similarly found, through qualitative evidence from 20 ISPO-certified smallholder institutions, that certified smallholders recorded higher average production than non-certified farmers. This improvement was attributed to the adoption of good plantation practices, which acted as a catalyst for higher FFB yields.

Extending this discussion, Rahutomo et al. [12] and Sugianto et al. [13] further showed that ISPO-and RSPO-certified smallholder institutions placed strong emphasis on Good Agricultural Practices, particularly crop maintenance, as a key mechanism for improving land productivity. Arifandy et al. [14] also found that the implementation of ISPO certification at the Sungai Galuh palm oil mill in Riau Province was followed by an increase in crude palm oil FFB production. In addition, Sugianto et al. [13] showed that improved plantation management practices among independent smallholders in several Indonesian provinces could substantially increase oil palm FFB and farmer profits.

As limited empirical literature found associated with ISPO, the broader literature on international certification schemes, particularly the RSPO, also provides relevant evidence on the potential production benefits of sustainability certification. Brandi et al. [15] found that RSPO certification generated economic benefits for smallholders in Sumatra through improved harvest performance, reflected in both the number and total weight of FFB. Morgans et al. [16] similarly reported that the proportion of RSPO-certified concession areas in Kalimantan was positively associated with FFB production. Renner et al. [17, 18] found that RSPO certification was positively correlated with land productivity and farm profits among smallholders in Jambi.

De Vos et al. [19] also found that RSPO-certified smallholders in Central Kalimantan tended to achieve higher FFB yields than non-certified farmers across most tree-age categories. At the same time, the literature provides some cautionary findings. Veriasa et al. [20], using district-level panel data from Riau during 2012–2021, found that RSPO-certified plantation area had a positive but statistically insignificant effect on oil palm productivity. Shahimi et al. [21] further showed that certification does not necessarily improve financial performance, particularly when compliance costs are high and market incentives are limited. Purnama et al. [22] also emphasized that certification should be viewed as one component of a broader sustainability strategy, requiring technological innovation, improved production practices, and stronger market incentives.

Several studies further suggest that the effects of certification are shaped by adoption constraints, institutional readiness, and market incentives. Adiprasetyo et al. [23] examined the determinants of smallholder participation in ISPO certification in Bengkulu, while Reich and Musshoff [24] investigated smallholders’ willingness to participate in certification programs in Jambi. Watts et al. [25] identified weak land legality and institutional capacity as major barriers to certification adoption in Central Kalimantan. De Vos et al. [9] similarly emphasized that pre-existing farmer organization and clear land legality improve the likelihood of successful certification.

Apriani et al. [26] found that financial incentives, access to credit, better market access, and improved FFB production outcomes motivated smallholders to maintain RSPO certification, while Hutabarat [27] noted that traditional plantation management, weak farmer groups, weak land legality, and poor documentation practices continue to constrain certification adaptation among Indonesian smallholders. Majid et al. [6], in a systematic review of RSPO, MSPO, and ISPO studies, concluded that empirical evidence on the direct economic and productivity effects of ISPO remains limited, although farmer organizations associated with certification may generate positive outcomes. Taken together, these studies suggest that certification may be positively associated with palm oil FFB production. Therefore, we develop the following hypothesis:

H1: Sustainable palm oil certification has a positive association with palm oil FFB production, and the effect persists across districts under different quartiles.

2.2 Fertilizer use and oil palm production

Fertilizer use is one of the most important agronomic inputs in oil palm cultivation because oil palm is a nutrient-demanding crop whose productivity depends heavily on adequate and timely nutrient application. Several studies indicate that low fertilizer use remains a major constraint on smallholder oil palm productivity in Indonesia. De Vos et al. [28] emphasized that adequate fertilizer application and proper seed management are essential for improving productivity. De Vos et al. [28] similarly found that low fertilizer use among Indonesian oil palm smallholders contributed to lower FFB production per hectare. Their findings further indicate that limited fertilizer application is associated with financial constraints, logistical barriers, poor infrastructure, and inadequate fertilizer supply. These constraints may prevent smallholders from applying fertilizer at the recommended level, thereby reducing their potential yield.

Other evidence reinforces the importance of fertilizer use in supporting palm oil FFB production. Research in Bengkulu Province shows that fertilizer use, particularly organic fertilizer, has a positive effect on oil palm production [29]. In addition, plantation area, fertilizer use, and institutional arrangements have been identified as important supporting factors for improving FFB production [30]. Renner et al. [17] also found that RSPO certification was associated not only with higher land productivity and farm profits but also with greater NPK fertilizer use, suggesting that certification may partly affect production through improved input application. This is consistent with the broader certification literature, which suggests that productivity gains are more likely when fertilizer use is embedded within good agricultural practices [12, 13, 19].

Accordingly, districts with a higher share of agricultural businesses using fertilizer are expected to record higher oil palm production. Fertilizer use may reflect both direct nutrient application and broader adoption of improved plantation practices. Based on this reasoning, we propose the following hypothesis:

H2: Fertilizer use is positively associated with oil palm production.

2.3 Farmer groups and oil palm production

Farmer groups are important institutional vehicles for improving oil palm production, particularly among smallholders who often face constraints in accessing knowledge, inputs, finance, extension services, and certification programs. The literature on certification adoption suggests that farmer organization is a key condition for successful participation in sustainability schemes. De Vos et al. [19] argued that certified smallholders often had stronger pre-existing group organization before certification, implying that part of the observed certification advantage may reflect prior institutional capacity. De Vos et al. [9] similarly found that pre-existing farmer organization and clear land legality improved the likelihood of successful certification and better plantation management. Majid et al. [6] also noted that farmer organizations associated with ISPO implementation may generate positive outcomes, suggesting that institutional arrangements can serve as a channel through which certification affects farm-level performance.

Several studies on certification adoption further highlight the role of farmer groups and collective organization. Adiprasetyo et al. [23] examined the factors influencing smallholder participation in ISPO certification in Bengkulu, while Reich and Musshoff [24] investigated the willingness of independent smallholders in Jambi to participate in certification programs. Watts et al. [25] found that weak institutional capacity was a major barrier to certification adoption in Central Kalimantan. Apriani et al. [26] showed that support programs facilitated by non-governmental organizations were important for helping smallholders maintain RSPO certification, while Hutabarat [27] identified the limited effectiveness of smallholder groups as one of the major constraints in adapting RSPO certification in Indonesia. These findings suggest that farmer groups may improve production indirectly by facilitating access to extension services, certification assistance, input markets, knowledge transfer, and collective learning.

Farmer groups may also strengthen palm oil FFB production performance by improving coordination among farmers, supporting the dissemination of good agricultural practices, and enhancing access to government or non-government support programs. Younger and less experienced farmers may also be more receptive to new knowledge and improved plantation management practices, which can be transmitted through farmer groups and extension networks [26]. Therefore, districts with more farmer groups are expected to have stronger palm oil FFB production capacity. Based on this argument, we propose:

H3: Farmer groups promote performance in oil palm production.

2.4 Institutional capacity, legal entities, and oil palm production

Institutional capacity and legal business status are central to improving oil palm production because they influence access to finance, formal markets, certification programs, technical assistance, and regulatory compliance. Several studies show that weak institutional structures and unclear legal status constrain certification adoption and production improvement among smallholders. Watts et al. [25] identified the lack of formal land ownership documentation and weak institutional capacity as major barriers to sustainability certification in Central Kalimantan. De Vos et al. [9] similarly emphasized that clear land legality and pre-existing farmer organization improve the likelihood of successful certification and better farm management. These findings suggest that formalization and institutional readiness are important foundations for improving production outcomes.

Widyatmoko [31] argued that ISPO certification can help address agrarian conflicts in oil palm plantations while accelerating the formalization of land ownership among smallholder producers. Such formalization may improve access to credit, markets, and institutional support, thereby strengthening production capacity. Majid et al. [6] also suggested that institutional arrangements associated with ISPO implementation may generate positive effects, even though direct evidence on ISPO’s productivity impact remains limited. Purnama et al. [22] further argued that certification initiatives such as ISPO need to be supported by technological innovation, better production practices, and stronger market incentives. This implies that certification and production improvement require not only technical compliance but also stronger institutional and business structures.

The broader literature also supports the importance of institutional and organizational arrangements. Apriani et al. [26] found that access to credit, market access, and financial incentives were important motivations for smallholders to maintain certification, while Hutabarat [27] noted that weak land legality and poor documentation practices constrain certification adaptation. Sugianto et al. [13] and Salmiyati et al. [30] further emphasized that good plantation management and institutional arrangements are important supporting factors for improving FFB production. In this study, the number of farming businesses with legal entities is used to capture the formal institutional and business dimension of oil palm production at the district level. Legally registered farming businesses are likely to have better access to finance, stronger managerial capacity, clearer administrative documentation, and greater ability to engage with formal markets and certification systems. Therefore, we propose:

H4: Agriculture businesses with legal entities are positively associated with oil palm production.

Table 2 presents two descriptive statistics, Model 1 and Model 2 for the study variables in their original units of measurement, prior to logarithmic transformation. In Model 1, average palm oil FFB production is 122,724.3 tons, with a relatively large standard deviation of 166,695 tons. This wide dispersion suggests considerable heterogeneity in production capacity across districts. The minimum recorded production is only 60 tons, while the maximum reaches 798,460 tons, indicating that the sample includes both marginal producing districts and major palm oil FFB production centers. Similarly, the average ISPO-certified area is 44,363.82 hectares, with a standard deviation of 54,363.97 hectares and a maximum value of 312,993.40 hectares. This pattern suggests that ISPO-certified plantation areas are unevenly distributed across districts, with certification concentrated in a relatively small number of large producing areas.

Table 2. Descriptive statistics

Source: The authors.

The control variables also show meaningful cross-district variation. On average, 57.4 percent of agricultural businesses use fertilizer, although the ratio ranges from 0.051 to 0.98, indicating that fertilizer adoption within agricultural sector differs substantially across districts. The average number of subdistricts is 14.272, with a minimum of 4 and a maximum of 51, reflecting differences in administrative and territorial scale. The number of beginner-level farmer groups is also relatively high, with an average of 1,214 groups per district, but the large standard deviation of 783.56 indicates considerable variation in farmer organization across districts. Meanwhile, the number of agricultural businesses with legal status shows extreme dispersion, with an average of 190.83 but a maximum of 21,205. This suggests that formal agricultural business institutions are highly concentrated in only a few districts.

We also provide information of descriptive statistics in Model 2 (using 197 sample size), as it alternatively measures ISPO certification in dummy format (1 if district has ISPO certification area with no less than 3 ha, both for company-based and smallholder). The mean value of the company-based ISPO certification dummy is 0.152, indicating that only about 15.2 percent of districts have more than the specified threshold of ISPO-certified company-managed plantation area. By contrast, the mean value of the smallholder-level ISPO certification dummy is 0.635, suggesting that approximately 63 percent of districts have ISPO-certified independent smallholder areas. This contrast implies that smallholder-level certification is more widely distributed across districts than company-level certification. Overall, the descriptive statistics highlight substantial heterogeneity in FFB production, certification coverage, fertilizer use, farmer organization, and formal institutional capacity, justifying the use of econometric models that account for differences across districts and production distributions.

To briefly inform palm oil country-level data as reported by USDA [33] to provincial level, Figure 1 illustrates the 2023 provincial-level summary of oil palm FFB production in Indonesia. It shows that Riau Province recorded the highest production, with approximately 478 thousand tons, followed by Jambi with 143 thousand tons, South Sumatra with 112 thousand tons, and Bengkulu with around 99 thousand tons. These figures indicate the dominant role of Sumatra as Indonesia’s main center of palm oil FFB plantation production.

1.png

Figure 1. Palm oil fresh fruit bunch (FFB) production in 2023 by province

Figure 2 shows a different pattern in terms of ISPO-certified oil palm area. Districts in East Kalimantan Province account for the largest ISPO certified oil palm area, while districts in Riau Province, despite having the highest FFB production level, rank second in terms of accumulated ISPO-certified area. This descriptive evidence suggests that higher production is not necessarily linearly associated with a larger accumulated ISPO-certified area. Therefore, while certification may be positively related to FFB production performance, descriptive patterns alone indicate that the relationship is likely shaped by other factors, including plantation scale, productivity, management quality, institutional capacity, and regional production characteristics.

2.png

Figure 2. Palm oil area certified by Indonesian Sustainable Palm Oil (ISPO) (in ha) by province

The scatter plot (Figure 3(a)) shows a positive association between ISPO-certified area and palm oil FFB production, both measured in logarithmic form. The fitted regression line slopes upward, indicating that districts with larger ISPO-certified oil palm areas tend to record higher levels of palm oil FFB production. The clustering of observations around higher ISPO-certified areas also indicates that certification is more commonly observed in districts with relatively established production systems. Furthermore, the bar chart in Figure 3(b) indicates a clear descriptive difference in average production between districts with and without company-level ISPO-certified areas. Districts without company-certified ISPO areas record an average log production value of 8.46, whereas districts with company-certified ISPO areas show a substantially higher average value of 11.97. This pattern tends to inform that higher production of palm oil FFB in districts with larger concentration of ISPO-certified area organized by company, though we cannot sure as the figure provided only for descriptive analysis.

3.png

3b.png

Moving to Table 3 which reports the OLS estimates of the association between ISPO-certified oil palm area and district-level oil palm FFB production. The results show a consistently positive and statistically significant coefficient for ISPO-certified area across all specifications. In Column (1), the coefficient is 0.3514 and statistically significant at the 1 percent level, indicating that a 1 percent increase in ISPO-certified oil palm area is associated with an approximately 0.351 percent increase in oil palm FFB production, ceteris paribus. Comparatively, after the full set of control variables is included in Column (5), the coefficient declines to 0.2974 but remains statistically significant at the 1 percent level. This suggests that the positive association between ISPO certification and palm oil FFB production is robust, although part of the initial relationship is explained by fertilizer use, farmer organization, administrative scale, and farmer business with legal status.

Table 3. Ordinary least squares (OLS) estimate of the effects of certified Indonesian Sustainable Palm Oil (ISPO) on palm oil fresh fruit bunch (FFB) production

Note: Robust standard errors in parentheses *** p < 0.01, ** p < 0.05, * p < 0.1.

The control variables also provide important insights into the determinants of palm oil FFB production. The ratio of agricultural farmer using fertilizer is positive across all specifications and remains statistically significant in the full model, although only at the range between 5 to 10 percent level. Since the dependent variable is expressed in logarithmic form and fertilizer use is measured as a ratio, the coefficient represents a semi-elasticity. In Column (5), for instance, a one-percentage-point increase in the share of agricultural businesses using fertilizer is associated with an approximately 1.79 percent increase in palm oil FFB production, holding other variables constant. Overall, the explanatory power of the model improves substantially as additional covariates are included, with the R-squared increasing from 0.1151 in Column (1) to 0.3647 in Column (5). This indicates that the full specification explains approximately 36.47 percent of the variation in district-level FFB production.

Table 4 reports OLS estimates controlling province fixed effects. Compared with Table 3, the coefficient of ISPO-certified oil palm area still remains positive and statistically significant at the 1 percent level across all specifications. However, its magnitude is slightly smaller after province fixed effects are included. In the full specification, the coefficient declines from 0.2974 in Table 3 to 0.2575 in Table 4. The reduction in coefficient size suggests that part of the relationship observed in Table 3 was explained by differences across provinces. Nevertheless, the persistence of a positive and highly significant coefficient confirms that the association between ISPO certification and palm oil FFB production remains robust after accounting for provincial heterogeneity.

Table 4. Ordinary least squares (OLS) estimates controlling province fixed-effect of the effects of certified Indonesian Sustainable Palm Oil (ISPO) on palm oil fresh fruit bunch (FFB) production

Note: Robust standard errors in parentheses *** p < 0.01, ** p < 0.05, * p < 0.1.

The inclusion of province fixed effects also changes the behavior of several control variables. The fertilizer-use ratio remains positive, but its statistical significance disappears in the fuller specifications, unlike in Table 3, where it remains significant at the 5-10 percent level in the full model. This suggests that the effect of fertilizer use may partly reflect province-level differences in input access, infrastructure, or agricultural support systems. The number of beginner-level farmer groups remains positive and statistically significant, although its coefficient declines from 1.1450 in Table 3 to 0.6301 in Table 3, indicating that farmer organization continues to be associated with higher FFB production but with a more moderate estimated effect after we control for province characteristics. Finally, the explanatory power of the model increases substantially, with the full specification R-squared rising from 0.3647 in Table 3 to 0.7303 in Table 4. This sharp improvement indicates that province-level factors explain a large share of the variation in district-level palm oil FFB production, while the main finding on ISPO certification remains stable and robust.

Table 5 reports the estimated association between ISPO certification (measured as dummy variables where if district has ISPO certification is given 1 if area is at least 3 ha, otherwise 0) and logged palm oil FFB production. Across all specifications, both districts with smallholder-level or company-level ISPO certification dummies are positive and statistically significant at the 1 percent level. This indicates that districts with ISPO-certified smallholder areas and districts with ISPO-certified company-managed areas tend to record substantially higher oil palm fruit yield than districts without such certification. The results are robust across the three estimation approaches: baseline OLS, OLS with province fixed effects, and OLS with clustered standard errors. The inclusion of province fixed effects reduces the estimated coefficients of both ISPO dummy variables, from 1.4217 to 1.1934 for smallholder certification and from 2.8595 to 2.4883 for company certification. This decline suggests that part of the observed association is explained by province-level characteristics. The clustered specification also confirms the statistical robustness of the main ISPO variables.

The magnitude of the coefficient is larger for company-level certification than for smallholder-level certification, suggesting that ISPO-certified company-managed plantations are more strongly associated with higher FFB production outcomes. Note here as dependent variable is expressed in logarithmic form, the percentage impact interpretation is, for example for Dummy_area certif_company coefficient in column (2), 100 ×[exp(2.4883)-1] = 1,104.1% higher than districts without having any ISPO certification, holding all else constant.

Table 5. Comparing impact of Indonesian Sustainable Palm Oil (ISPO) as dummy variables on palm oil fresh fruit bunch (FFB) production

Note: Robust standard errors in parentheses *** p < 0.01, ** p < 0.05, * p < 0.1.

To summarize, Table 5 provides strong evidence that ISPO certification, particularly company-level certification, is positively associated with logged oil palm FFB, while institutional formalization also appears to be an important supporting factor.

Finally, Table 6 presents the quantile regression estimates for oil palm FFB production across the 25th percentile, median, and 75th percentile of the FFB production distribution. Columns (1)–(3) use the logarithm of ISPO-certified oil palm area as the main explanatory variable, while Columns (4)–(6) use dummy variables for districts with ISPO-certified smallholder and company-managed plantation areas, respectively. The results indicate that ISPO-certified area is positively associated with palm oil FFB production across all quantiles. The coefficient is 0.3527 at the 25th percentile, 0.3542 at the median, and 0.1560 at the 75th percentile. This suggests that the association between ISPO-certified area and FFB production is stronger among lower-and middle-producing districts than among districts in the upper part of the FFB production distribution. The effect is positive across quantiles, although the evidence of Model 1 at the 75th quantile is weaker and only marginally significant at the 10% level.

Table 6. Quantile regression

Note: Bootstrapped standard errors in parentheses, based on 500 replication, *** p < 0.01, ** p < 0.05, * p < 0.1.

The dummy-variable specifications in Columns (4)–(6) further support the positive association between ISPO certification and palm oil FFB production. The dummy for district with ISPO-certified smallholder areas is positive and statistically significant across all quantiles, with the strongest coefficient observed at the median. This indicates that districts with ISPO-certified smallholder areas tend to record higher palm oil FFB production than districts without such certification, particularly around the central part of the FFB production distribution. The dummy for district with ISPO-certified company-managed areas is also positive and statistically significant at the 1 percent level across all quantiles, with the largest coefficient at the 25th percentile. This pattern suggests that company-level ISPO certification is strongly associated with higher FFB production, especially among lower-producing districts. However, because the dependent variable is in logarithmic form and the dummy coefficients are relatively large, these estimates should be interpreted primarily as log-point differences and as strong conditional associations rather than precise causal percentage effects.

The control variables show heterogeneous patterns across quantiles and model specifications. In Columns (1)–(3), the fertilizer-use ratio is positive and statistically significant at the 25th percentile and median, but not at the 75th percentile. Since fertilizer use is measured as a ratio while the dependent variable is in logarithmic form, its coefficient represents a semi-elasticity. The results imply that fertilizer use is more relevant for FFB production among lower- and middle-producing districts, whereas its independent association is weaker among higher-producing districts. The number of beginner-level farmer groups is also positive and statistically significant across all quantiles in Columns (1)–(3), suggesting that farmer organization may support FFB production through collective learning, extension access, and better adoption of plantation practices. However, its statistical significance weakens in the dummy-variable specifications, indicating that part of the farmer-group effect may overlap with certification status or other institutional characteristics.

The number of subdistricts has a negative coefficient in most specifications and is statistically significant mainly at the 25th percentile, suggesting that administrative or territorial dispersion may be more strongly associated with lower FFB production among districts in the lower segment of the FFB production distribution. By contrast, the number of agricultural businesses with legal status is positive across all specifications and becomes statistically significant particularly in the upper quantile of the continuous ISPO model and across all quantiles in the dummy-variable specifications. This finding reinforces the importance of formal institutional and business structures in supporting palm oil FFB production. Overall, the quantile regression results confirm that ISPO certification is positively associated with palm oil FFB production, but the magnitude of this association varies across the production distribution. The higher pseudo R-squared values in the dummy-variable specifications suggest greater explanatory power, although this comparison should be made cautiously because the number of observations differs between the continuous-area and dummy-based models.