THE CALCULATION OF RATES OF RETURN ON AGRICULTURAL RESEARCH FROM PRODUCTION FUNCTIONS

A general framework is described whereby, using production functions, the rate of return on additional agricultural R & D can be calculated. This enables existing procedures to be critically reviewed. Illustrative results are given which indicate that the rates of return found in previous studies tend to be overestimates. It is also pointed out that these studies usually refer to well in the past, when the level of the agricultural research investment was generally lower and rates of return consequently larger than might now be expected. It is, however, concluded that the production-function approach needs extension and refinement, and better values of key parameters determined, before it could be used quantitatively as a determinant of research policy.     

AN INDIVIDUAL TRAVEL-COST METHOD OF EVALUATING FOREST RECREATION

Consumer surplus for outdoor recreation has traditionally been estimated by a Clawson-Knetsch travel-cost method. This paper presents zonal consumer-surplus estimates for visitors to a number of forests and compares these estimates to those derived from individual visitor observations. Both travel-cost procedures are used to assess the magnitude of recreational benefits and are found to produce widely different consumer-surplus estimates. This raises questions about research methodology and has implications for the value of recreation associated with forestry and its contribution to the rate of return on forest investment.     

Canadian dairy policy and the returns to federal dairy cattle research

The economic surplus approach is used to estimate the returns to federal investments in dairy cattle research in Canada. A national supply function is estimated using time series data. Lagged research expenditures are included as explanatory variables in the model, facilitating the calculation of ***marginal as well as average benefits from research. Simulation analysis is used to study the effects of product market distortions associated with Canadian dairy policy as well as of the marginal excess burden on the rates of return to research and on the distribution of research benefits. Returns were found to be high at the margin. Distortions in the product market had a small effect on the overall returns to dairy cattle research but had a large impact on the distribution of research benefits. Rate of return estimates were found to be indicative of underinvestment even when the marginal excess burden was taken into account.     

Decision-making Criteria for Purchasing Milk Quota at the Individual Farm Level

Dairy farmers wishing to contemplate purchases of additional quota should first consider what the appropriate time horizon on their investment should be. If the chosen time horizon is one year or less, the problem of estimating how much they can afford to bid for quota can be solved with simple budgeting techniques. On the other hand, if the time horizon is taken as several years, and this would seem rational, then capital budgeting techniques are required. However, in either case, investment in additional quota should be compared with alternative investment opportunities before a final decision is made. Where capital budgeting techniques are used, a modification of the net present value approach permits the farmer to take account of his time preference for money, and allows calculation of the break-even price, or the maximum affordable price to pay for additional quota. This approach necessitates the provision of estimates of the expected net returns from investing in quota for each year of the selected time horizon, as well as provision of estimates of the salvage value of the quota and an appropriate discount rate. The calculation of the expected net returns figures represents the most difficult task for the farmer. If expansion of milk shipments rests upon increasing output of milk per cow, then the short-run production function and marginal cost curve provide the appropriate reference points. The expected net return on investment in quota is then given by the area under the marginal revenue line and above the relevant portion of the marginal cost curve. If, however, all short-run variable input decisions have already been implemented at the optimum level, and expansion of milk shipments depends upon expanding the number of cows, or the “fixed” plant and equipment, or both, then the average total cost curve is the appropriate reference point. The expected net return on investment in quota is then represented by the difference between total net revenue at the new planned level of output and total net revenue at the existing level of output.     

Quasi‐maximum likelihood estimation of a censored equation system with a copula approach: meat consumption by U.S. individuals

A copula approach to censored system estimation is proposed. The quasi‐maximum likelihood estimator departs from the multivariate normal error distribution predominantly used in existing estimators and resolves the computational difficulty with multiple probability integrals in high‐dimensional censored systems. An application to individual meat consumption demonstrates that the procedure produces very different empirical estimates from existing Gaussian full‐information and quasi‐maximum likelihood estimates.     

Short and long‐run returns to agricultural R&D in South Africa,or will the real rate of return please stand up?

This paper briefly presents the results of a total factor productivity (TFP) study of South African commercial agriculture, for 1947‐1997, and illustrates some potential pitfalls in rate of return to research (ROR) calculations. The lag between R&D and TFP is analyzed and found to be only 9 years, with a pronounced negative skew, reflecting the adaptive focus of the South African system. The two‐stage approach gives a massive ROR of 170%. The predetermined lag parameters are then used in modeling the knowledge stock, to refine the estimates of the ROR from short‐ and long‐run dual profit functions. In the short run, with the capital inputs treated as fixed, the ROR is a more reasonable 44%. In the long run, with adjustment of the capital stocks, it rises to 113%, which would reflect the fact that new technology is embodied in the capital items. However, the long‐run model raises a new problem since capital stock adjustment takes 11 years, 2 years longer than the lag between R&D and TFP. If this is assumed to be the correct lag, the ROR falls to 58%, a best estimate. The paper draws attention to the possible sensitivity of rate of return calculations to assumed lag structure, particularly when the lag between changes in R&D and TFP is skewed.     

Recalibrating the reported returns to agricultural R&D: what if we all heeded Griliches?*

Zvi Griliches’ seminal analysis of hybrid corn spawned a large literature seeking to quantify and demonstrate the value of agricultural research and development (R&D) investments. The most important metric for quantifying the rate of return to R&D emerging from this literature is the internal rate of return (IRR), even though Griliches was sceptical of its usefulness as a metric in this context. An alternative metric, also reported by Griliches but not as commonly used in the subsequent returns‐to‐research literature, is the benefit–cost ratio (BCR). We assess how the implications of the returns to agricultural R&D literature may have differed if the BCR had become the standard rather than the IRR. We reveal that the IRR and BCR produce substantially different rankings of agricultural R&D projects, differences that persist even under various commodity and geographical aggregations of the BCR and IRR estimates. The median across 2,627 reported IRRs is 37.5 per cent per year. Using data gleaned from 492 research evaluation studies, we developed and deployed a methodology to impute 2,126 BCRs (median of 5.4) and modified internal rates of returns (MIRRs, median of 16.4 per cent per year) assuming a uniform 10 per cent per year discount rate and a 30 year research timeline.     

The Calculation of returns to research in distorted markets: Comment

In a recent article, Oehmke reported that a high internal rate of return for investment in research when the interaction between research and price policy costs was disregarded could become very low or even negative when the effects of research on the costs of price policy were considered. In this paper, the social returns from research in the presence of the price policies considered by Oehmke are reexamined using a simple geometric approach. The analysis suggests that an output subsidy in a small importing economy, an output subsidy in a closed economy, and a target price in a large exporting economy will on Oehmke's assumptions - cause only small reductions in the internal rate of return from investment in research. This implies that the apparent underinvestment by governments in agricultural research cannot be explained away by a large upward bias, known to governments, in measured rates of return due to failure to account for interactions between research and the costs of price policy measures.     

The Economic impact of public agricultural research and development in the United States

In this article, we examine the relationship between public investments in agricultural research and development and the productivity‐enhancing benefits they generate. Knowledge productivity functions are estimated for U.S. agriculture using data on multifactor productivity and public knowledge stocks. We examine the time‐series properties of the data and compare alternative econometric estimation procedures. The results are used to calculate economic performance measures such as internal rates of return and benefit‐cost ratios. The real rate of return to public investments in agricultural research and development in the United States is in the range of 8–10% per annum.     

Measuring public agricultural research and extension and estimating their impacts on agricultural productivity: new insights from U.S. evidence

This article provides new estimates of the marginal product of public agricultural research and extension on state agricultural productivity for the U.S., using updated data and definitions, and forecasts of future agricultural productivity growth by state. The underlying rationale for a number of important decisions that underlie the data used in cost‐return estimates for public agricultural research and extension are presented. The parameters of the state productivity model are estimated from a panel of contiguous U.S. 48 states from 1970 to 2004. Public research and extension are shown to be substitutes rather than complements. The econometric model of state agricultural TFP predicts growth rates of TFP for two‐thirds of states that is less than the past trend rate. The results and data indicate a real social rate of return to public investments in agricultural research of 67% and to agricultural extension of 100+%. The article concludes with guidance for TFP analyses in other countries.     

Land use planning and the ecosystem approach: An evaluation of case study planning frameworks against the Malawi Principles

Global land use change continues to undermine the capacity of ecosystems to sustain ecosystem service (ES) flows. Much attention in policy and research has therefore been given to concepts, tools and processes for sustainable land use planning, including consideration of ES and the ecosystem approach. However, there are limited empirical cases or evaluations of ecosystem approach based planning from which lessons can be drawn. The aim of this research therefore was to identify and evaluate existing case study planning frameworks that have the potential to operationalise the ecosystem approach. Based on the Malawi Principles, a new suite of evaluation criteria was developed. This was used to assess case study documentary evidence and evaluate the extent to which the 12 Malawi Principles had been considered. The evaluation also assessed the planning methods/approaches used by the case studies and their potential to help translate the Malawi Principles into land use planning outcomes. Finally, a SWOT analysis was used to structure the main findings. Our results show that the Malawi Principles have been considered across the case studies “fully” or “partially” in 64% of instances suggesting, therefore, that the case studies present a reasonable interpretation of the ecosystem approach. However, poor consideration of biodiversity and environmental limits across the cases highlights the risk of land use management decisions continuing to contribute to the degradation of natural capital.     

The Calculation of Returns to Research in Distorted Markets

Most of the empirical literature calculating rates of return to publicly sponsored research assumes that research is the only relevant government intervention. For most countries this assumption is untenable. This paper shows that improperly measuring government induced market distortions can severely bias research rate of return calculations. If the interaction between successful research and other government interventions increases the cost of the other interventions, then neglecting market distortions unambigously increases the calculated rate of return. Three examples of government induced distortions show that the magnitude of the upward bias in calculated rates of return can be extremely large - in some cases more than 100 percentage points. A normative implication is that governments should account for interactions between research and price interventions when determining research support levels. A positive implication is that existing government research funding patterns are more readily explainable as reasonable behavior by a government that accounts for these interactions.     

Returns from research in economies with policy distortions: hybrid sorghum in Sudan

Conventional estimates of the economic return to agricultural research use market prices for the values of products and inputs; on this basis, economic rates of return are typically well above the cost of capital, suggesting that more investment in research would be socially desirable. But these estimates may be incorrect if, as is often the case, market prices are distorted by market failures or government policies and hence do not reflect social values. This paper presents a simple, partial-equilibrium methodology with which to improve the measurement of social returns to research by taking account of multiple distortions in the market prices of products, inputs and foreign exchange. The method also takes account of variation in domestic and world prices, making a product tradable in some years and nontradable in others. The method is applied to the case of Hageen-Dura 1 (HD-1), Sub-Saharan Africa's first commercially successful hybrid sorghum. HD-1 was released in the Sudan in 1983. From the start of research in 1979 to 1992, the HD-1 breeding program had an estimated IRR of 97% when all major policies in the sorghum market, the fertilizer market, and the exchange rate are taken into account. The high rate of return to HD-1 research was due to the program's low cost and rapid payoff, pointing to the potential value of small adaptive research programs, taking full advantage of foreign technology and genetic material to produce locally-appropriate crosses in a short period of time. Even in the highly distorted economies of Africa, such programs can yield very high payoffs.     

Research returns redux: a meta-analysis of the returns to agricultural R&D

A total of 289 studies of returns to agricultural R&D were compiled and these provide 1821 estimates of rates of return. After removing statistical outliers and incomplete observations, across the remaining 1128 observations the estimated annual rates of return averaged 65 per cent overall — 80 per cent for research only, 80 per cent for extension only, and 47 per cent for research and extension combined. These averages reveal little meaningful information from a large body of literature, which provides rate-of-return estimates that are often not directly comparable. This study was aimed at trying to account for the differences. Several features of the methods used by research evaluators matter, in particular assumptions about lag lengths and the nature of the research-induced supply shift.     

Nonparametric Estimation of Oligopsony Power in First‐Price Auction

This paper proposes a nonparametric procedure to estimate market power for first‐price auction data and applies the procedure to cattle procurement markets. Most previous studies have used parametric methods, which require specific functional forms for retail demand, input supply and processors’ cost equations. However, researchers often find that market power estimates from parametric methods are sensitive to the choice of functional forms and specifications. An application to data from cattle procurement experiments shows that our nonparametric approach greatly outperforms commonly used parametric methods in estimating the degree of market power. While parametric estimates are sensitive to functional form specification and are at least 90% smaller than ‘true’ market power indices, estimates from our nonparametric procedure deviate from the actual value by no more than 25%.     

Linking technical change to research effort: an examination of aggregation and spillovers effects

We used a disaggregate approach to examine investment efficiency of wheat breeding research in India. India's total research effort comprizes 20 research programs spread across 50 experiment stations. A technology spillover matrix was constructed for both potential and actual spillovers. Spillovers and free‐riding were dominant characteristics of technical change during the period studied. Although the aggregate rate of return to wheat improvement research in India was estimated to be 55%, eight programs were found to have earned a negative rate of return when spillins were taken into account. Research output is concentrated on a few strong programs. The two strongest programs generated 75% of all the technical change benefits, even though they claimed just 22% of research resources. These two programs include a significant degree of overlap, while on the other hand many farmers were not reached by any of the programs – 56 and 78% of rainfed and durum area, respectively, in 1990 was still sown with pre‐1976 varieties.     

THE RETURNS TO INVESTMENT IN RESEARCH ON AUSTRALIAN WOOL PRODUCTION*

A production function approach is used to estimate growth in farm productivity in the Australian wool industry from an estimated level of expenditure on wool production R & D. A market equilibrium model of the wool industry is then used to measure the share of total benefits from this productivity growth accruing to Australia and its wool growers. A net return is estimated after allowing for lags in the development and adoption of technology.     

Productivity and the Returns to Levy-Funded R&D for Sugar Production in the Eastern Counties of England

Estimates of the rate of return (ROR) to publicly funded-agricultural research are getting lower as private expenditures and spill-overs are more adequately handled. For UK sugar beet there is a pool of technology available and the spill-ins are not measurable. An alternative approach is to assume that the difference between productivity growth in sugar and the rest of UK agriculture is attributable to the Sugar Beet Research and Education Committee's R&D and extension expenditures, funded by the only long-standing producer levy in the UK. These expenditures are used to explain the difference between total factor productivity (TFP) growth in sugar (3.5 per cent per annum) and the rest of UK agriculture (2.0 per cent per annum). The producer's ROR calculated using this approach is 11 per cent and the lower bound on the total return, to producers and consumers is 21 per cent, whereas the conventional methodology gives returns of 87 per cent. Thus, the upward bias in ROR calculations may be removed by changing the approach to the problem.     

Returns to public sector plant breeding in the presence of spill‐ins and private goods: the case of bean research in Michigan

This article illustrates a methodology for assessing economic returns to a publicly funded breeding program in the presence of private sector investments, and spill‐ins from other contemporary public institutions and past research efforts. The approach consists of determining yield gains from bean improvement research; applying these yield gain estimates to measure benefits attributable to different institutional players and time periods; and then assessing the benefit‐cost ratios of investments in a bean improvement program since 1980 by Michigan State University (MSU). The results indicate that investments in MSU's bean breeding program have yielded benefits to costs ratio in the range of 0.7 to 2.2, depending on the attribution rule used to estimate the benefits. The estimated benefit/cost ratios reported in this study are lower‐bound estimates, as they do not account for potential benefits from area planted to MSU varieties outside of Michigan (spillover effects), which was 1.5 times greater than the area planted to MSU‐bred varieties within Michigan in the period 1998–2002. The implications of the increasingly privatized bean seed markets for the role of public sector research in bean improvement research are discussed.     

Ex post impacts of improved maize varieties on poverty in rural Ethiopia

Public agricultural research has been conducted in Africa for decades. While many studies have examined its aggregate impacts, few have investigated how it affects the poor. This paper helps fill this gap by applying a new procedure to explore the ex post impacts of improved maize varieties on poverty in rural Ethiopia. Plot‐level yield and cost changes due to adoption are first estimated using instrumental variable and marginal treatment effect techniques where possible heterogeneity is carefully accounted for. A backward derivation procedure is then developed to link treatment effect estimates with an economic surplus model to identify the counterfactual household income that would have existed without improved maize varieties. Poverty impacts are finally estimated by exploiting the differences between observed and counterfactual income distributions. Improved maize varieties have led to a 0.8–1.3 percentage drop of poverty headcount ratio and relative reductions of poverty depth and severity. However, poor producers benefit the least from adoption due to the smallness of their land holdings.