Valuing the impacts of agricultural research and development investments

Australia’s agricultural productivity challenge is becoming more complex, requiring innovation that manages uncertainty, coordinates investment, and builds long-term knowledge. While cost–benefit analysis (CBA) remains essential for demonstrating R&D value, it systematically undervalues research that shapes future decisions or reduces uncertainty. At the same time, the accumulation of CBAs represents an underused evidence base for portfolio-level insight. This article argues for extending, not replacing, current frameworks by integrating Value of Information and stronger portfolio analysis to better guide investment in productivity, resilience, and competitiveness.

Innovation and productivity

Australia's agricultural productivity has been fundamentally shaped by public and private investment in research and development (R&D). Innovation enabled by research has played a critical role in supporting sectoral growth, improving sustainability, responding to biosecurity threats, and maintaining the competitiveness of Australian producers in global markets.

For producers contributing levies, governments allocating funding, and researchers proposing future projects, a foundational question must be answered: what value does this investment deliver?

The answer is rarely straightforward. Outcomes are diffuse, long-term, and highly variable across regions and production systems, and the tools used to measure them have not kept pace with the complexity of modern agricultural R&D portfolios.

Gross agricultural production is forecast to reach $101 billion in 2025–26, a record high.^[1] Around 70% of that output is exported, making Australian agriculture one of the most trade-exposed and globally competitive sectors in the economy, and one of the most important to national income, regional employment and food security.

But the headline figures mask a structural challenge that is intensifying. Since the early 2000s, long-run annual broadacre farm productivity growth has slowed to around 0.7% per year, less than a third of the 2.2% average achieved during the 1980s and 1990s.^[2] While favourable seasons and technological adoption sustain some momentum, climate variability is an increasing drag, and the "easy" productivity gains from the previous generation of innovation have largely been captured. The task now requires harder, more complex, and more coordinated investment to deliver equivalent results.

This matters through supply chains and regional economies. Agriculture is one of the few sectors of the Australian economy where productivity has continued to outpace the national average. When it slows, the costs are distributed widely through supply chains, regional economies, export earnings, and ultimately the national budget.

The question for policy makers, Rural Research and Development Corporations (RDCs), levy payers and researchers is not whether to invest in R&D, but how to invest better, and how to demonstrate that the investment is working.

The RDC system: a proven co-investment model

Australia's RDC system is the backbone of public agricultural innovation. Established in the late 1980s, the system now comprises 15 RDCs spanning many agriculture, fishery and forestry sectors. 

Together, RDCs translate about $1.3 billion per year in levy funds, Commonwealth matching contributions and co-investment into practical innovation across Australia's agricultural sector.^[3] Industry levies collected from producers are matched by the Australian Government up to a statutory cap. Research is commissioned in response to strategic priorities and delivered through universities, CSIRO, state agencies, and an increasingly significant private sector.

The returns are well established. ABARES estimates that every $1 invested in agricultural R&D generates an almost $8 return to farmers over a 10-year period.^[4] A cross-RDC aggregation of investments from 2014 to 2019, conducted using Council of Rural RDCs-aligned methodology, found a weighted average benefit-to-cost ratio of 4.4 over a 30-year horizon, in other words, roughly $4.40 in benefits for every dollar invested.^[5]

These are strong numbers. The challenge is understanding what they capture, and what they miss.

Cost-benefit analysis: essential, but not sufficient 

Cost-benefit analysis (CBA) remains a primary and robust tool for valuing agricultural R&D. It translates invested dollars into measurable returns, provides a consistent basis for comparing investments, and offers accountability to levy payers and taxpayers.

The Council of Rural RDCs Impact Assessment Guidelines standardise the approach across the RDCs, mandating consistent discount rates, time horizons, and treatment of non-market benefits such as environmental improvements and biosecurity risk reduction.

ACIL Allen has applied this methodology across numerous R&D impact assessments, including work for the Grains Research and Development Corporation, Hort Innovation, Australian Pork Limited, Sugar Research Australia, and others. We found that CBA remains the appropriate foundation for demonstrating accountability and guiding portfolio allocation. The framework works especially well where impacts are direct, adoption is measurable, and the causal chain from investment to outcome is clear. But rigidly applied, project-by-project CBA has structural blind spots that become more problematic as the productivity task grows harder.

The first is the adoption bottleneck. Realising benefits from R&D requires producers to adopt new practices, technologies or systems: a process that is slow, uneven and shaped by factors well beyond the quality of the research itself.^[6]  

When these factors are reduced to a single assumption in a BCR calculation, they introduce significant uncertainty, and can disadvantage long-term, high-reward innovations that require structural on-farm change. 

The second is the attribution problem. Agricultural outcomes are the product of multiple simultaneous investments, policy settings, market conditions and on-farm decisions. Attributing a measurable share of a yield improvement or disease reduction to a single RDC project requires expert judgement, scenario analysis and producer surveys.^[7] The "hard numbers" in many assessments rest on assumptions that are softer than they appear.

The third, and most important, is that CBA applied in isolation cannot adequately value research that produces knowledge rather than an immediately adoptable technology.

The value of what research eliminates

In a competitive R&D portfolio, a significant proportion of projects will not produce commercial technology. Some will demonstrate that a pathway is infeasible. Some will narrow the landscape of viable options for subsequent, more targeted investment. Some will establish the evidence base that allows industry bodies to avoid committing resources to a dead-end track. Under a standard project-level BCR, these outcomes look like failures. In reality, they are a critical function of a well-managed portfolio.

For example, GRDC's investment in establishing a detailed greenhouse gas emissions baseline for the Australian grains sector undertaken with CSIRO from 2020 did not directly produce an adoptable on-farm product.[8] But it established the foundational data that enabled the subsequent Low Emissions Intensity Farming Systems (LEIFS) initiative, which now equips growers with practical tools to monitor, manage and report farm-scale emissions.[9] Without the baseline, the applied program could not have been designed with confidence.

Similarly, Sugar Research Australia's investment in NIR spectroscopy for Ratoon Stunting Disease detection is built on a long sequence of prior investment in diagnostic science and disease surveillance. The commercial-scale screening system now being integrated into sugar mills is the product of an innovation pipeline, not a single project.

This is not a failure of the traditional framework. CBA is designed to measure the value of outcomes that can be defined. It is less well suited to valuing research that reshapes future decisions, research that changes what questions the industry needs to ask next.

One highly useful framework in this context is the Bayesian Value of Information (VoI). This method quantifies the economic value of reducing uncertainty in strategic decision-making.

1. Example A (Avoided Costs): A research project investigates whether an expensive new pest management strategy is cost-effective. If the study proves it is not, and this prevents the industry from sinking investments into an unpromising pathway, the research has delivered immense value by avoiding wasted resources.
2. Example B (Strategic Foundation): A project maps the long-term, emerging risks of climate change on pasture productivity. Even if this data does not change daily on-farm practices immediately, it is critical for guiding future adaptation research and helping industry bodies proactively structure their funding priorities.

Evaluating each project in isolation will systematically undervalue research that is foundational, exploratory, or enabling. A more accurate picture requires tracing how each project contributes to a broader sequence of research, recognising that the study which ruled out a dead end was not a failure, but a necessary step toward the one that succeeded.

Impact assessment practice must evolve

Agricultural R&D is not a pipeline of projects delivering incremental gains. It is a system for navigating uncertainty in an increasingly constrained and volatile production environment. If evaluation frameworks continue to privilege what is easiest to measure, they will quietly steer investment away from the research that matters most.

This is an argument to make fuller use of cost–benefit analysis, not to move beyond it. The Council of RDCs Impact Assessment Guidelines already provide the methodological consistency that makes this possible. Because CBAs across the RDC system are conducted using aligned discount rates, time horizons and treatment of non-market benefits, the accumulated body of assessments constitutes a coherent longitudinal dataset, not simply a collection of individual project appraisals. Applied consistently over time and across commodities, priorities and organisations, this dataset can identify where returns are strongest, where adoption constraints persist, and where the system is over- or under-investing. The value of CBA is therefore cumulative: any single assessment is partial, but together they enable portfolio-level learning and more strategic allocation of public and levy funds.

Realising that potential requires moving from treating CBAs as end-point accountability tools to treating them as inputs into a broader evidence framework. Integrating Value of Information analysis alongside conventional CBA would allow the system to recognise the full economic role of research: not only as a generator of outputs, but as a guide to better decisions under uncertainty. The central task is to evolve the existing approach, understanding portfolios rather than isolated projects, valuing long-term capability alongside short-term returns, and treating uncertainty not as a modelling inconvenience but as the central problem that agricultural R&D exists to solve.