The US Conservation Reserve Program: the evolution of an enrollment mechanism Daniel Hellerstein, Economic Research Service, USDA Presented at the Conservation Tenders in Developed and Developing Countries - Status Quo, Challenges and Prospects workshop Jakobsberg, Boppard, Germany, Sept 12 2013 1
Key points The CRP is over 25 years old, and is shrinking in size The CRP relies on an auction mechanism that incorporates a ranking mechanism (the EBI) and a maximum bid (a bid cap, aka the SRR) The bid cap may be problematic Alternative auction mechanisms may be more effective 2
What is the Conservation Reserve Program (CRP) Established in 1985. Focus is on retiring cropland As of July 2013, about 11m ha (27m ac) enrolled in 10 to 15 year contracts; at total cost of around $1.6 billion/year. 8.7m ha in general signup, 2.2 m ha in continuous signup Originally targeted at highly erodible land. Since 1990s, a broader array of concerns Soil erosion Water quality Air Quality Wildlife habitat Enduring benefits Carbon sequestration Slide 3 of 24
Two types of CRP Land is entered into CRP in two ways: 1. General signup Competitive enrollment Most land enters the CRP this way 2. Continuous signup Non-competitive enrollment Many contracts, less land (smaller contracts on average) Focus on priority land 4
CRP 2013 acreage: general General signup acres 1 dot = 800 ha Approximate centroid of general signup acreage 5
CRP 2013 acreage: continuous Continuous signup acres 1 dot = 800 ha Approximate centroid of continuous signup acreage 6
CRP 2013 acreage: general and continuous 1 dot = 800 ha Brown: continuous Green: general Approximate centroid of all acreage 7
Recent trends in CRP enrollment: shrinking size, with continuous enrollment a growing share 16 35% decline since 2007! Million ha 14 12 10 8 6 4 2 0 Source: CRP contract data (FSA/USDA) New & re-enrolled general New & re-enrolled continuous Existing general Existing continuous 8
The CRP the operational environment Optional participation Land costs, and environmental benefits, are heterogeneous. Landowners can increase the environmental values of offered lands (quality improvement) The government has limited information on the opportunity cost of the land (noisy assessments) Slide 9 of 24
The CRP s general signup enrollment mechanism Several week enrollment period every year Parcels must meet eligibility criteria: cropland history and either erodibility or location (in a Conservation Priority Area). Landowners submit offers that specify land cover and a bid Results are announced several weeks later.. 10
Key features of the general signup Offers are sealed bid Accepted offers are paid their requested rental rate All offers are ranked using an Environmental Benefits Index (EBI) The EBI incorporates the requested rental rate The requested rental rate can not exceed a parcel specific maximum price. This bid cap is called the Soil Rental Rate (SRR) 11
The General Signup Environmental Benefits Index (EBI), Circa 2012 100 150 100 45 50 100 Wildlife Water quality Erosion reduction Enduring benefits Air quality Cost savings Note: the EBI is not a benefit/cost ratio. Instead, cost enters as an additive factor. 12
2012 County average bid caps (SRR) 2012 Acreage Mean: $51 Median: $39 10%ile $25 90%ile $97 $ per acre Source: FSA data 13
Soil rental rates and commodity prices (relative to 1997) Prices 2,5 2 1,5 1 0,5 0 corn price wheat price US average SRR Commodity prices and the CRP s SRR were fairly stable from 1997 to 2006 Since 2007, commodity prices have been substantially higher Since 2007, the CRP s SRR have increased somewhat Source: Average SRR computed from USDA/FSA offer files. Commodity prices from http://www.indexmundi.com/commodities/ 14
Trends in general signup offers Number of acres offered varies over time, as does the acceptance rate ha (offered & accepted) 10 9 8 7 6 5 4 3 2 1 0,9 0,8 1 0,1 0 0 1997 1998 1999 2000 2003 2006 2010 2011 2012 2013 ha offered ha accepted Accept rate 0,7 0,6 0,5 0,4 0,3 0,2 Acceptance Rates Source: CRP contract and offer data (FSA/USDA) 15
... bids tend to be close to their bid cap Bid as fraction of SRR 1 0,98 0,96 0,94 0,92 0,9 0,88 0,86 0,84 0,82 Differences between bids and SRR, by year and SRR 22 27 33 40 49 60 73 90 109 134 164 200 244 SRR (the bid cap) Source: Linear regression using CRP contract data (FSA/USDA) 1998 Lower productivity lands tend to bid their maximum, 16
... bids tend to be close to their bid cap Bid as fraction of SRR 1 0,98 0,96 0,94 0,92 0,9 0,88 0,86 0,84 0,82 Differences between bids and SRR, by year and SRR 22 27 33 40 49 60 73 90 109 134 164 200 244 SRR (the bid cap) Source: Linear regression using CRP contract data (FSA/USDA) 1998 2004 Lower productivity lands tend to bid their maximum, Over time, bids across all parcels are closer to bid caps 17
... bids tend to be close to their bid cap Bid as fraction of SRR 1 0,98 0,96 0,94 0,92 0,9 0,88 0,86 0,84 0,82 Differences between bids and SRR, by year and SRR 22 27 33 40 49 60 73 90 109 134 164 200 244 SRR (the bid cap) Source: Linear regression using CRP contract data (FSA/USDA) 1998 2012 2004 Lower productivity lands tend to bid their maximum, Over time, bids across all parcels are closer to bid caps 18
Parcel specific bid caps and cost heterogeneity If a single price were paid to all offers, owners of low cost parcels could earn substantial rents A precise bid cap (equal to a parcel s opportunity cost) could deliver substantial savings However, a poorly chosen bid cap can increase total expenditures Slide 19 of 24
Simple Example: bid cap issues Assume a population of 10 parcels with opportunity costs ranging from 1 to 10 Goal is to enroll 5 parcels Type of offer Total cost Actual value 15 Single price (6 th highest value) 30 Using an upwardly biased assessment (+2% to +5% greater than actual) Using accurate & unbiased assessment (+/- 2% of actual ) 18.7 31.5 20
Lab results -- Impacts of different bid caps 4 types of tickets; tickets of the same type have the same bid cap Variable Coefficient (T-Stat) Intercept 4.6 (9.4) 90% Bid cap 90% of type maximum -0.13 (-29.3) 100% -0.25 (-2.9) 120% -0.064 (-1.4) 200% Bid cap 200% of type maximum 0.0019 (0.03) Y = aggregate efficiency : values closer to 1 are better Coefficients can be interpreted as fractional changes compared to a baseline treatment of an 80% bid cap (i.e.; -0.10 = 10% decrease in over all cost) 21
Alternative mechanisms Are there enrollment mechanisms that would use available information more effectively that would replace bid caps with auction design? We considered two alternatives Quota auctions Reference price auctions 22
Quota Description A two stage acceptance procedure: Example of a score calculation (lower scores are better) Classify all offers into groups. For example, a low, medium, and a high soil productivity group. 1. Within each group, the worst scores are dropped 2. The survivors are pooled, re ranked, and the best offers are accepted Score= BID q (q=quality enhancement purchased) The same bid, hence same, score is used in both stages 23
Reference Price Description A reference price is assigned to each parcel. This could be the SRR. Bids below the reference Price: better score Bids above the reference Price: worse score Offers are ranked by score, and the best scoring offers are accepted Example of a score calculation scores are better) Score= BID - 0.5q + (BID referenceprice) The impact of two identical bids will depend on the referenceprice 24
Experimental results: the impact of quota and reference price mechanisms Variable Aggregate efficiency (cost of acquisition / full information cost) Dependent variable Average Profit Reference Price -0.18 (-8.9) -10.5 (-12.9) Quota -0.14 (-5.81) -7.7 (- 8.3) maxprior 0.0054 (6.8) 0.21 (7.1) R-square [ f-stat ] 0.66 [< 0.001] 0.58 [< 0.001] Negative values on the Reference Price and Quota dummies imply these mechanisms are more cost effective. 25
Summary The CRP has evolved over time Currently a mix of competitive enrollment (general signup) and targeted (continuous signup) Is the competitive ranking mechanism (the general signup) effective? Experimental evidence, and recent history, suggest a need to rethink the bid cap Alternative auction mechanisms may improve performance Is continuous, with its explicit targeting, better? Slide 26 of 24
With acreage likely to decrease, and with tight budgets, changes in the CRP are likely. Growing interest throughout the government in evidence based policy, which could lead to changes in program design Improved data and models on environmental performance could mean better targeting with a goal of landscape impacts rather than parcel specific metrics. Personally, I foresee increasing willingness of USDA to try new methods, which means interesting possibilities for experimental work (both lab and field) The end 27
Extra slides 28
The CRP is not the USDA s first land retirement program 90.0 CRP enrollment and other diverted acreage, 1982-2002 80.0 70.0 60.0 50.0 40.0 30.0 20.0 10.0 0.0 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 Year Set Set aside aside Total Total CRP CRP New New CRP CRP Set asides were mostly for production control, and were not managed for conservation purposes. Source: Land use data (FSA/USDA) 4
7,0 6,0 5,0 The CRP is USDA s largest conservation program (though not as largest as it once was) Trends in USDA Conservation Expenditures, 1983-2011* Ag Land Preservation (FPP and GRP) Working Land Conservation (EQIP and predecessors, CSP, WHIP, and others) Land Retirement (CRP and WRP) Conservation Technical Assistance (CTA) Billion $ 4,0 3,0 2,0 1,0 0,0 1983 1988 1993 1998 2003 2008 *Constant (2010) dollars Source: http://ers.usda.gov/topics/natural-resources-environment/conservation-programs/background.aspx 5
At the peak of the CRP-Soil Bank program in 1960, 28.7m acres were under contract. 31
CRP 1982-2009 enrollment, rental payments, and notable events Million acres 40 Billion dollars 2,5 35 30 2 25 20 15 10 5 Acres (millions) 1,5 1 0,5 0 0 CRP acres Continuous Annual program outlays 32
A multitude of program design decisions affect program outcomes Program Budget Program Performance All Farmers All Eligible Farmers Program Applicants Program Participants Economic Environmental Distributional Eligibility usually broad Land types Land use Location Practices Participation Incentive What action? Fields Practices What payment? Fixed Bid Enrollment Screen (targeting) Benefit-cost index Field Location Soils Topography Population Practices Cost (bid)
General signup acreage declining since 2007 Change in CRP acreage: 2007 to 2013 Dec 2012 crp acreage 1 dot = 1000acres = 404ha. ~35% acreage decrease in general signup acreage since 2007, with greater proportional decreases in the plains. Source: CRP contract data (FSA/USDA) 34
Change in CRP acreage from 2007 to 2013 : overall changes Counties where continuous and general acres are: Million acres % acreage # counties % counties bothdown both down 3.02 11 589 22 bothup both up 0.49 2 61 2 contdown continuous decrease > general increase 0.04 1 65 2 contup continuous increase > general decrease 2.16 8 445 16 gendown general decrease > continuous increase 21.14 78 1443 53 genup general increase > continuous decrease 0.14 1 27 1 nochanges nochanges 0.01 1 99 4 35
Maximum available EBI points in CRP signup #39 (2010) Wildlife [100] = Water quality [100] = Cover (introduced grass, native grass, trees) [50] + Wildlife enhancement [20] + Wildlife priority zones [30] Location within designated State water quality zone [30] + Groundwater quality [25] + Surface-water quality [45] Erosion [100] = Erodibility index [100] Enduring benefits [50] = Air quality [45] = Costs [150] = Source: FSA; see their program fact sheet for more details. Enduring benefits (tree plantings, wetland restoration, existing trees, grass seeding) [50] Wind erosion impacts [25] + Wind erosion soils [5] + In air quality zones [5] + Carbon sequestration [10] Per-acre rent [125] + Bid below maximum rate [25] 36
Location of CRP Conservation Priority Areas (CPA). 2010 37
Simple Example: bid cap issues Assume a population of Tidbits: 10 parcels with heterogeneous values (but otherwise the same) Type of offer Goal: accept 5 of the 10 parcels, by making parcel specific offers Total cost Actual value 15 Actual value Accurate and unbiased assessment Less accurate/upwardly biased assessment Single price (6 th highest value) Using less accurate assessment (upwardly biased +2 to +5%) Using accurate & unbiased assessment (+/- 2%) 30 18.7 31.5 38
.. expiring general signup parcels tend to be re-offered, and sometimes are the dominant set of offered acres Rates (offer and acceptance) 0,90 0,80 0,70 0,60 0,50 0,40 0,30 0,20 0,10 0,00 1997 1998 1999 2000 2003 2006 2010 2011 2012 2013 % of all offered acres that are "expiring" % of "expiring acres" that were re- offered Source: CRP contract and offer data (FSA/USDA) 39
Variable Coefficient T-Stat Intercept 4.6 9.4 90% Bid cap 90% of type maximum 100% 120% Lab results -- Impacts of different bid caps 4 types of tickets; tickets of the same type have the same bid cap 200% Bid cap 200% of type maximum Q10 Q20 Q40-0.13-0.25-0.064 0.0019 0.08-0.065-0.17-29.3-2.9-1.4 0.03 1.2-1.4-2.9 200% x Q40-0.13-1.6 Several other parameters are not displayed N F-stat Rsquare 183 (all treatments) 44.5 (prob<0.0001) 0.88 Y = aggregate efficiency = (observed quality adjusted cost ) / (full information cost for same quality) Lower values mean greater cost effectiveness. Coefficients on dummies can be interpreted as fractional changes (i.e.; 0.10 = 10%) 40
Experiments: quota and mechanism Using two ticket types, imposing a quota reduced acquisition costs by an average of 8%. Cost reduction due to reduced bids by low costs tickets was greater than cost increases due to accepting higher cost tickets 41
Experimental results: quota and reference price enrollment mechanisms 50 40 30 20 10 0 Cost and profit (experimental $) 60 Average results within a round for different treatments Standard Reference price Quota Cost Profit Profit rate 42
Continuous signup has many variants As of July 2013, the 2.2m ha of continuous signup CRP includes: 0.13m ha in the Farmable Wetlands Program 0.51 in 46 CREP agreements in 33 states 0.67 in several initiatives Continuous initiatives (maximum acres allocated) First Year Cumulative (Acres) Flood-plain wetlands (600,000 ac.) 2004 230,722 Bottomland hardwood trees (250,000 ac.) 2004 84,573 Non-flood plain and playa wetlands (350,000 ac.) 2005 224,804 Upland bird habitat buffers (500,000 ac.) 2005 241,678 Longleaf pine plantings (250,000 ac.) 2007 117,066 Duck nesting habitat (300,000 ac.) 2007 204,259 State acres for wildlife enhancement (1,250,000 ac.) 2008 726,142 Highly erodible lands (750,000 ac.) 2012 56,817 Pollinator habitat (100,000 ac.) 2012 826
CRP acreage: general and continuous 1 dot = 800 ha Brown: continuous Green: general 44
Continuous signup Continuous CRP is focused on environmentally sensitive land, with more stringent eligibility requirements. Buffers for Wildlife Habitat Wetlands Buffer Riparian Buffer Wetland Restoration Filter Strips Grass Waterways Shelter Belts Living Snow Fences 45 Contour Grass Strips Salt Tolerant Vegetation Shallow water areas for wildlife Source: USDA/FSA http://www.fsa.usda.gov/fsa/webapp?area=home&subject=copr&topic=crp
Continuous acreage is more targeted Eligibility for continuous signup is determined at the local level (for example, proximity to a waterway). Enrollment is non-competitive: if you are eligible, you will be automatically accepted Since various bonuses are available, payment rates under continuous signups are higher than under general signups (for the same parcel of land) A number of Initiatives address particular issues such as wildlife enhancement, flood plain wetlands, and pollinator habitat The Conservation Reserve Enhancement Program (CREP) is a subset of continuous, that comprises 1.3 million acres spread over 45 joint State and Federal programs 46
General vs continuous A limited comparison of the effectiveness of general and continuous signup (results of unpublished research from University of Illinois) Average Sediment Abatement (ton/acre) Average Nitrogen Abatement (kg/acre) 4 20 2 10 0 CREP IL CRP IL CREP MN CRP MN 0 CREP IL CRP IL CREP MN CRP MN Cost of Enrollment per Acre Sediment Abatement per Dollar $ per acre 200 150 100 50 0 I L La Moine CREP Spoon Vermilli CRP Cotton M N Redwood Watano Kg Abated per Dollar 6 4 2 0 La Moine Spoon CREP Vermilli CRP Cotton Redwood Watano