A cross-sectional study published in Frontiers in Agronomy examines conditional associations between firm characteristics and constructed monthly monetary operating surplus among agribusinesses in Chocó, Colombia. The research, conducted by Jose Fabian Rios Obando, Eddy Lorena Burbano Vallejo, and Jesús David Ponce Mendoza, provides exploratory firm-level evidence from a region with limited business registers and data. The study explicitly states that its observational, non-probability design neither identifies causal effects nor supports population-level inference, focusing instead on describing relationships within the specific sample of surveyed firms.
The investigation was motivated by the structural challenges facing Chocó, which remains the least competitive department in Colombia, consistently ranking at the bottom of the State Innovation Index. Notable areas of underperformance include innovation, institutional quality, and social conditions, mirroring national statistics that identify Chocó as the most impoverished region in the country. Despite policy commitments to sustainable economic growth and competitiveness, the department faces persistent infrastructure, financing, and formalization challenges that constrain agribusiness development.
In October 2024, researchers conducted a cross-sectional survey of 328 agribusiness owners across the department’s five subregions: Atrato, Darién, North Pacific, South Pacific, and San Juan. Due to the high degree of informality in business operations and the absence of comprehensive records regarding the total number of agribusinesses, a non-probability snowball sampling technique was employed. This method is appropriate for accessing hard-to-reach populations when a formal sampling frame is unavailable, though it may overrepresent connected enterprises and does not justify statistical generalization to the entire population.
The reported ordinary least squares (OLS) model included 301 complete observations, with 27 questionnaires excluded due to missing values in required variables. The outcome variable was the natural logarithm of monthly revenue less reported labor, transportation, and raw-material costs, calculated in Colombian pesos. This constructed monetary operating surplus serves as a narrow indicator of business performance, distinct from net accounting profit or multidimensional competitiveness, as other expenses such as taxes, depreciation, and financing costs were not included in the construction.
Higher employee education showed a positive conditional association with operating surplus in the reported model. The coefficient for the proportion of employees with higher education was 0.10, with a reported z-statistic of 3.50 and a p-value less than 0.001. If this measure is coded as a proportion from zero to one, a 0.10-unit increase corresponds to an approximately 1% higher constructed operating surplus, conditional on the included regressors. This descriptive association does not establish that education or training causes improved firm performance, as omitted firm characteristics and reverse causation remain plausible explanations.
The technology-investment coefficient was positive but not statistically significant based on the reported coefficient and standard error. The coefficient was 0.016 with a standard error of 0.08, implying a z-statistic of 0.20 and a two-sided p-value of approximately 0.841. The available figures do not support a statistically significant conditional association with constructed operating surplus in the reported specification. This result does not establish that technology investment is unimportant, but indicates insufficient statistical evidence in this specific model to claim a significant relationship.
Negative coefficients were observed for the labor, transportation, and raw-material cost ratios. The transportation-cost ratio had a coefficient of −2.12 (p = 0.009), the labor-cost ratio had a coefficient of −1.63 (p = 0.012), and the raw-material-cost ratio had a coefficient of −1.65 (p = 0.029). These associations are partly mechanical because the cost components are subtracted from revenue when calculating the constructed operating surplus. Therefore, these coefficients do not establish independent causal effects or demonstrate that specific cost policies cause changes in firm performance.
The model also included variables for land registration, supplier count, and collaborative deals, none of which showed statistically significant associations at the conventional 5% level. The cadastral registration coefficient was −0.051 (p = 0.617), and the supplier count coefficient was 0.0057 (p = 0.887). Statistical nonsignificance does not establish that these factors have no economic relevance or that associated policies are ineffective; rather, it indicates that the present regression does not identify their causal effects. Further research would be needed to examine selection into associations and their potential benefits.
Regional indicators in the model described conditional differences within the sample, with coefficients for Atrato and San Juan showing statistical significance at the 10% and 5% levels, respectively. However, the original coding and reference constraints for these categorical terms are not specified in the available summary. Consequently, the coefficients should not be interpreted as comparisons with a named omitted group or as population-wide regional effects. The inclusion of these terms does not eliminate unobserved confounding or support causal identification of governance quality or infrastructure effects.
The study utilizes heteroskedasticity-robust HC0 standard errors to address potential heteroskedasticity in the model. While these adjustments improve the reliability of inference regarding the error process, they do not remove omitted-variable bias, reverse causation, or measurement error. Firm size and ownership structure were not included in the published specification, and their omission may generate bias if these unmeasured characteristics are associated with both the included regressors and operating surplus. The analysis is therefore interpreted as descriptive and conditional on the variables included in the model.
The findings provide exploratory evidence limited to the analyzed sample, highlighting the need for further research using more comprehensive data and appropriate research designs. The non-probability sampling design, missing observations, and possible endogeneity preclude causal conclusions and population-representative generalizations. Future studies should employ probability sampling, panel designs, or credible causal identification strategies to better assess these relationships. The current analysis contributes to the understanding of firm-level dynamics in an underdocumented territorial setting without establishing the independent importance of structural constraints.
Policy implications are presented as considerations for further investigation rather than effects established by this regression. Recommendations regarding infrastructure, technology, and institutional support are informed by the regional context and cited literature, not by identified policy effects in this model. The study underscores the complexity of agribusiness competitiveness in peripheral regions, where territorial determinants such as infrastructure, education, and institutional coordination interact with firm-level capabilities to shape business performance.
This research contributes to the broader literature on agribusiness competitiveness by integrating operational, human-capital, technological, organizational, and regional measures in a single empirical framework. It demonstrates the importance of carefully distinguishing between constructed performance indicators and multidimensional competitiveness concepts. The study’s limitations highlight the challenges of conducting firm-level research in regions with high informality and limited data infrastructure, offering a foundation for more rigorous future investigations into the drivers of agricultural business success in Colombia.
Subject of Research: Conditional associations between firm characteristics and constructed monthly monetary operating surplus among agribusinesses in Chocó, Colombia.
Article Title: Infrastructure, capital, and labor: understanding agribusiness competitiveness in Chocó, Colombia
Article References: Rios Obando, J. F., Burbano Vallejo, E. L., & Ponce Mendoza, J. D. (2026). Infrastructure, capital, and labor: understanding agribusiness competitiveness in Chocó, Colombia. Frontiers in Agronomy, 8, Article 1957561. https://doi.org/10.3389/fagro.2026.1957561
Image Credits: AI Generated
DOI: 10.3389/fagro.2026.1957561
Keywords: agribusiness, Chocó, Colombia, operating surplus, employee education, cross-sectional study, OLS regression, non-probability sampling, firm performance, territorial development, Infrastructure, capital
News Source: Courtney Benton. (November 8, 2026). Study Links Employee Education to Agribusiness Surplus in Chocó. Scienmag.



