Substitutes versus Complements among Canadian Business Risk Management Programs.

Transcription

1 Substitutes versus Complements among Canadian Business Risk Management Programs Nicoleta Uzea a*, Kenneth Poon b*, David Sparling c, Alfons Weersink d a Post-Doctoral Research Associate, Ivey Business School, Western University, nuzea@ivey.ca b Research Associate, Department of Food, Agricultural and Resource Economics, University of Guelph, poonk@uoguelph.ca c Professor and Chair of Agri-Food Innovation, Ivey Business School, Western University, dsparling@ivey.ca d Professor, Department of Food, Agricultural and Resource Economics, University of Guelph, aweersin@uoguelph.ca *Senior authorship is shared by the first two authors. Selected Paper prepared for presentation at the Agricultural and Applied Economics Association s 2014 Annual Meeting, Minneapolis, MN, July 27-29, Copyright 2014 by [Nicoleta Uzea, Kenneth Poon, David Sparling and Alfons Weersink]. All rights reserved. Readers may make verbatim copies of this document for non-commercial purposes by any means, provided that this copyright notice appears on all such copies. 0

2 INTRODUCTION Governments worldwide spend considerable amounts of money developing programs and policies that are meant to help their agri-food sector survive and grow in the increasingly competitive and volatile marketplace. In Canada, federal and provincial governments spent more than $33.6 billion 1 in total over the last agricultural policy framework, Growing Forward I ( ) that is 29.1% of the GDP generated in agriculture over the five-year period or 13.2% of the agri-food GDP (see Figure 1 for the annual statistics). As Figure 1 shows, the amounts spent under Growing Forward II ( ) are not likely to change substantially, at least based on preliminary figures for 2013 and budget estimates for Figure 1. Federal and Provincial Government Expenditures in Support of the Agri-Food Sector, Government Expenditures in Support of the Agri-Food Sector, % Growing Forward I $7.5B $6.7B $6.4B $6.4B $6.6B $6.6B p $6.2B e Growing Forward II Total Federal and Provincial Expenditures 30% 20% 10% Government Expenditures as % of Agriculture GDP Government Expenditures as % of Agri-Food GDP 0% p 2014e Source: Agriculture and Agri-Food Canada, Farm Income, Financial Conditions and Government Assistance Data Book, Notes: p = preliminary figures based on actuals and budget estimates when actuals are not available; e=estimated figures based on budget estimates. Given the significant amounts of public money that governments spend on programs, a number of questions bear consideration. For instance, do programs perform as intended? What can be done to improve their effectiveness? Are there better ways of allocating resources (across programs) so that the greatest possible benefits are obtained? This paper aims to answer the first two questions by examining the business risk management (BRM) programs the cornerstone of Canadian agricultural policy and to open up discussion on the third question. 1 52% of these expenses were incurred at the federal level, with the remaining 48% at the provincial level. 1

3 Share Business risk management (BRM) continues to be the central objective of Canadian agricultural policy even with the changes introduced under the Growing Forward II policy framework. As Figure 2 shows, farm support payments 2 make up for the largest portion of government expenditures in support of the agri-food sector over the period , the share of program payments varied between 48% in 2009 and 41% in By comparison, R&D programs attracted only between 1.8% of government expenditures in 2009 and 3.3% in 2013, while the share of marketing and trade programs varied between 1.1% in 2009 and 1.6% in Figure 2. Government Expenditures in Support of the Agri-Food Sector, by Major Category, Government Expenditures in Support of the Agri-Food Sector, by Major Category, % Growing Forward I Growing Forward II 40% Operating and Capital 30% Program Payments 20% 10% Research and Inspection Development and Trade Related Programs Others 0% p 2014e Source: Agriculture and Agri-Food Canada, Farm Income, Financial Conditions and Government Assistance Data Book, Notes: p = preliminary figures based on actuals and budget estimates when actuals are not available; e=estimated figures based on budget estimates. BRM programs are meant to provide protection against market volatility and disasters, while not impeding the flow of market signals and adaptation of the sector within the dynamic marketplace. Appendix 1 briefly describes the suites of Canadian BRM programs under the three most recent policy frameworks. While there have been changes in the way they operate, programs remained fairly consistent across the three policy frameworks and include: 1) AgriInsurance (and its predecessor Production Insurance); 2) AgriStability; 3) AgriInvest; 4) AgriRecovery, and 5) Advance Payments Program. The programs have been designed to work together in reducing business risk by providing protection against different types of risks and by allowing payments to be made at different times (Table 1). Specifically: 2 Include payments for Income Support and Stabilization programs, Ad Hoc and Cost Reduction programs, Crop/Production Insurance programs, and Financing Assistance programs. 2

4 - AgriInsurance provides coverage for production losses related to specific crops or commodities caused by natural perils; - AgriStability provides coverage for large declines in farm income caused by circumstances such as low commodity prices, rising input costs, and production losses; - AgriInvest helps cover small income shortfalls or make investments to reduce on-farm risks; - AgriRecovery helps producers return their farm businesses to operation following disaster situations by filling gaps not covered by the other BRM programs; - Advance Payments Program improves cash flow and provides producers with flexibility in marketing their commodities, thus allowing them to benefit from best market conditions. In terms of timing: - AgriInsurance payments are made soon after a loss, providing the needed cash flow during the production year; - AgriStability final payments are made after the production year is completed (interim payments are available to provide early access to partial benefit); - AgriInvest withdrawals can be made anytime throughout the year. Table 1. Business Risk Management Programs Risks Covered and Timing of Payments Program Risks covered Timing of payments AgriInsurance Production losses During production year, soon after a loss AgriStability AgriInvest Large income declines (due to low commodity prices, rising input costs & production losses) Small income shortfalls & investments to reduce on-farm risks After production year is completed (interim payments available within 30 days) Anytime throughout the year AgriRecovery Disaster situations Advance Payments Program Cash flow shortages Spring and fall In addition, a number of incentives have been built into the programs to encourage producers to use the programs together rather than substitute for one another. 3 For example, because production losses covered under AgriInsurance can also be covered under AgriStability and/or AgriInvest (provided they translate into a whole farm loss 4 ), some farmers may choose to participate in AgriStability and/or AgriInvest only. The following incentives are provided to encourage producers to also participate in AgriInsurance: 3 Note that AgriRecovery is not a program for which producers can choose to apply in advance once Governments decide that an initiative will be developed under AgriRecovery, the details of that initiative are made available to producers, including the process required to access the program. 4 It may be that increased prices cancel production losses; hence, AgriStability payments cannot be triggered, while the farm could benefit from AgriInsurance proceeds. This effect is going to get even stronger with the drop in AgriStability coverage to 70% under Growing Forward II. 3

5 a) AgriStability payments are reduced by the amount of losses that could have been covered under AgriInsurance; b) AgriInsurance payments are included as allowable income in a producer s Reference Margin under AgriStability and Allowable Net Sales under AgriInvest. Maintaining a higher Reference Margin means a higher level of income is protected under AgriStability. Also, the larger the Allowable Net Sales, the larger the amount a producer can deposit in the AgriInvest account and the larger the matching government contribution. However, AgriStability payments or government contributions under AgriInvest are not included as allowable income in the Reference Margin under AgriStability or the Allowable Net Sales under AgriInvest. c) AgriInsurance payments can help fully cover negative margins. AgriStability provides negative margin protection at 60% under Growing Forward I and 70% under Growing Forward II, and d) AgriInsurance provides coverage for spot losses. AgriInsurance losses are determined on a crop specific basis. If a farmer does not have a whole farm AgriStability loss, he/she may still be protected through AgriInsurance for one or more crops. Also, because producers may choose not to participate in AgriInsurance and/or AgriStability hoping that, if a disaster strikes, they would receive assistance under AgriRecovery, AgriRecovery is only offered when AgriInsurance and AgriStability cannot respond to the disaster. If it is determined that AgriStability and AgriInsurance coverage provides assistance to respond to the disaster, then additional coverage under AgriRecovery is not offered to those producers who chose not to participate in these programs. Finally, in order to qualify for a cash advance under the Advanced Payments Program, a farmer must participate in another business risk management program such as AgriInsurance. Despite such efforts to encourage producers to participate in the programs, anecdotal evidence and a few recent studies suggest that programs may crowd out each other. For instance, Kimura and Anton (2011) provide evidence based on simulation analysis that CAIS/AgriStability reduces farmers incentives to participate in Crop Insurance/AgriInsurance. Also, commodity prices may have implications for how AgriInvest and AgriStability are used. Specifically, an increase in prices increases the contribution room for AgriInvest (contribution is based on sales), hence, increasing the incentive to participate in AgriInvest; however, it decreases the likelihood that a farm can trigger AgriStability payments, hence, decreasing the incentive to participate in AgriStability. Similarly, a drop in prices will encourage participation in AgriStability and potentially discourage participation in AgriInvest (though the high rate of return on AgriInvest contributions is expected to mitigate this effect). Moreover, producer participation tends to drop in most programs, which raises questions about the relevance and performance of these programs. Available data at the Ontario level (Figure 3) shows that AgriStability and AgriInvest cover an increasingly smaller share of farms and market revenues. The exception is Crop Insurance, but increased participation in this program is likely due mostly to the cross compliance requirement for the Risk Management Program (RMP) imposed in

6 Figure 3. Participation Pattern in Ontario Sample Percentage of Farms vs. Percentage of Market Revenue, Program Participation as % of Farms vs. % of Market Revenue, % Ag Policy Framework Growing Forward I 90% 80% 70% 60% 50% 40% CAIS/AS part (% of farms) CAIS/AS part (% of market rev) CI part (% of farms) CI part (% of market rev) AI part (% of farms) AI part (% of market rev) The objective of this paper is four-fold: 1) understand the factors that influence producers decision to participate in the main BRM programs Crop Insurance, AgriStability, and AgriInvest; 2) understand the interlinkages between participation in these programs; 3) examine if and how those factors and relationships differ across farm sectors and sizes, and 4) assess the impact of program participation on the business risk experienced by producers. The answers to these questions are central for policy makers in order to achieve the desired objectives. If indeed there is substitutability between any of the programs, the current BRM suite of programs will need to be re-designed. Moreover, if programs are not effective at reducing business risk, the whole BRM policy will need to be re-evaluated. EMPIRICAL MODEL The analysis in this paper is conducted in two stages. In the first stage, we estimate dynamic multivariate and multinomial probit models with unobserved heterogeneity to study: 1) the interrelated dynamics of participation in Crop Insurance (CI), AgriStability (AS), and AgriInvest (AI) (participation in the programs is assumed to be endogenously determined), and 2) the factors that affect participation in each of the programs. In the second stage, we use the predicted probabilities for different participation states from the first stage to examine the impact of program participation on business risk. We estimate a business risk function using quantile regression and tobit models. 5

7 Program Participation Decisions To study the interrelated dynamics of participation in the programs, we first estimate a dynamic multivariate probit model such as the one specified in equations (1)-(6): = Ι if 0 and 0 else (4) = Ι if 0 and 0 else (5) = Ι if 0 and 0 else (6) For an individual i (subscript not shown for simplicity), the probability of participation in CI at time t is expressed in terms of a latent variable as specified in equation (1), the probability of participation in AS at t is expressed by the latent variable, specified in equation (2), and the probability of participation in AI at t is expressed by the latent variable, specified in equation (3). The dependent variables are the dummy indicators,, and specified in equations (4)-(6) they are equal to 1 if the individual participates in the program at time t and 0 otherwise. Explanatory variables are the lagged participation states (,, and ) and exogenous regressors ( ), including: enterprise diversification, crop diversification, profitability, operating efficiency, debt coverage, past business risk, and reliance on government payments.,, and represent the time-invariant error terms (unobserved heterogeneity) and and represent the time-specific idiosyncratic shocks. The consistent estimation of this model requires a solution to the initial conditions problem (Heckman, 1981). The problem arises due to our lack of knowledge of the data generating process that governs the start-of-period status for participation in CI, AS and AI. We use the approach suggested by Wooldridge (2005) to correct for the initial condition. As an alternative to the multivariate model presented above, we also estimate a dynamic multinomial probit model with unobserved heterogeneity in which the choice set is made up of all possible combinations of the three BRM programs instead of just the programs by themselves. Specifically, there are eight possible combinations that a producer can chose to adopt: 1) participate in all three programs (CI, AS, and AI) simultaneously; 2) participate in CI and AS; 3) participate in CI and AI; 4) participate in AS and AI; 5) participate in AS only; 6) participate in AI only; 7) participate in CI only, and 8) participate in none of the programs. Given this choice set, the dynamic multinomial probit model can be specified as follows: + + 6

8 where represents the program combination j (j = 1,, 8) that the ith producer chooses to participate in at time t and is the lagged participation state. The variable BPM is coded from 1 to 8, so that equation (7) is only one equation to be estimated. Combination 8 (i.e., no participation) is chosen as the base category and excluded from the estimation for identification purposes. The multinomial model allows us to examine how different variables affect the probability of choosing a certain combination of BRM programs, rather than how explanatory variables affect participation in one particular program, conditional on participation in the other programs. Impact of Program Participation on Farm Business Risk To assess the impact of participation in BRM programs on a farm s business risk, we estimate the following business risk function using both tobit and quantile regression: DR = γz + ε (8) The dependent variable, downside risk, is measured as the percentage change in gross margin from previous year. The independent variables include, among others, a set of 8 state dummies for all possible combinations of program participation. Since these participation states are latent and endogenous, they are replaced by predicted probabilities from the first stage. For instance, (1,1,1) is replaced by the probability that a farm participated in all three programs. Given that these propensities add up to one, it is necessary to use one combination as a reference category to avoid perfect collinearity; the (0,0,0) combination (i.e., no participation) is the reference category. The coefficients on these participation state dummy variables provide preliminary evidence on the complementarity/ substitutability of these programs in reducing business risk. That is, two programs are complementary if they mitigate downside risk more when used jointly than when either program is used in isolation. z is a vector of control variables including enterprise diversification, crop diversification, and size dummies. DATA Data Source The dataset used in the analysis resulted from merging three datasets: 1) the Ontario Farm Income Database (OFID); 2) AgriCorp s Crop Insurance Data, and 3) Agriculture and Agri-Food Canada (AAFC) s AgriInvest Data. Data merge was possible through a common identifier (Personal Identification Number). OFID is a longitudinal farm-level dataset including all Ontario tax-filing operations from 2003 to The dataset contains detailed financial, production and program payment (except for Crop Insurance) data. Additional operator-level Crop Insurance and AgriInvest payment data was linked to 7

9 farm-level OFID records to complement the program payment data. The panel sample size after cleaning 5 is 8,721 farms, distributed across sectors 6 as follows: Sector # of farms % of total sample Field Crops 5, % Beef % Supply Managed (Dairy, Poultry) % Swine % Others 1, % TOTAL 8,721 The analysis is conducted on the aggregate sample, as well as for field crops versus beef the two largest sectors in the sample. Variable Definition Variable Name Dependent Variables Crop Insurance AgriStability AgriInvest Downside Risk Independent Variables Sector diversification Crop diversification Operating Profit Margin Operating Expense Ratio Debt Coverage Ratio Payment Reliance Business Risk Farm Size Dummies $10,000-$100,000 (omitted) $100,000-$250,000 $250,000-$500,000 >$500,000 Description =1 if the producer participates in Crop Insurance at time t, 0 otherwise =1 if the producer participates in AgriStability at time t, 0 otherwise =1 if the producer participates in AgriInvest at time t, 0 otherwise Measured as the percentage change in gross margin from previous year Calculated as a Herfindahl index of revenue allocations among various operations (e.g., crops, beef, dairy): lower values=greater diversification Calculated as Herfindahl index of revenue allocations among various crops: lower values=greater diversification Measure of profitability. Calculated as net operating income (before interest and taxes)/ total operating revenue Measure of operating efficiency. Calculated as total operating expense/ total operating revenue Calculated as net operating income (before interest and taxes)/ interest expense Calculated as the share of program payments in total revenue Calculated as the coefficient of variation of revenue/expense of previous five years =1 if a farm has average total revenue between $10,000-$100,000 over the panel period, 0 otherwise =1 if a farm has average total revenue between $100,000-$250,000 over the panel period, 0 otherwise =1 if a farm has average total revenue between $250,000-$500,000 over the panel period, 0 otherwise =1 if a farm has average total revenue > $500,000 over the panel period, 0 otherwise 5 Farms with average annual revenue less than $10,000 were dropped. 6 Sectors are defined based on share of revenues in six out of the nine years. 8

10 RESULTS A key result is that there is positive and statistically significant state dependence in program participation. As the transition probability matrix in Table 2 shows, producers who participate in a program or program combination in the current year tend to do so next year, and those who do not participate in any of the programs in the current year tend to not participate next year. The positive and significant coefficient of the lagged participation variables in the multivariate model (Table 3) supports this result, even after controlling for other factors and unobserved heterogeneity. Table 2. Transition Probabilities of Participation States in AgriStability (AS), AgriInvest (AI), and Crop Insurance (CI), Participation State in Period t+1 Participation AS=1 AS=1 AS=0 AS=1 AS=1 AS=0 AS=0 AS=0 State in AI=1 AI=0 AI=1 AI=1 AI=0 AI=1 AI=0 AI=0 Period t CI=1 CI=1 CI=1 CI=0 CI=0 CI=0 CI=1 CI=0 AS=1 AI= CI=1 AS=1 AI= CI=1 AS=0 AI= CI=1 AS=1 AI= CI=0 AS=1 AI= CI=0 AS=0 AI= CI=0 AS=0 AI= CI=1 AS=0 AI= CI=0 Also, we find that programs tend to be used together rather than independently the pairwise correlations between the residuals in the multivariate model (see rho in Table 3) are positive and significant, with the correlation between participation in Crop Insurance and AgriStability the strongest, followed by that between AgriStability and AgriInvest. Significance of these error correlations means we 9

11 can reject the hypothesis of independence independence of the two participation decisions, confirming the need for their joint estimation. As regards the factors that affect participation, we find that operating efficiency increases the propensity to participate in Crop Insurance and decreases the propensity to participate in AgriStability and AgriInvest. A potential explanation for why efficient farms are more likely to participate in Crop Insurance, but less likely to participate in AgriStability is that efficiency can compensate for any production losses hence, they may not be able to trigger AgriStability payments, but they can still trigger Crop Insurance payments. Also, we find that larger farms are more likely to participate, one reason for that being more availability of human resources to sign up in the program, and to prepare and submit the claim. It may also be because these farms find it is worth the time given the larger amount of money they receive. A surprising result is that historic income volatility has little effect on participation, suggesting the decision is a forward-looking one. Another surprising (and robust) result is that crop specialization is inversely related to propensity to participate in AgriStability and Crop Insurance. One potential explanation for the relationship between specialization and AgriStability is that farmers may use marketing contracts to protect against price risks (though we lack information on the use of marketing contracts). Note that the results from the multinomial model (Table 4) generally support these findings. 10

12 Table 3. Parameter Estimates from the Multivariate Probit Model for Estimating the Factors Affecting Participation in Crop Insurance, AgriStability, and AgriInvest All Farms Field Crops Beef Crop Insurance AgriInvest AgriStability Crop Insurance AgriInvest AgriStability Crop Insurance AgriInvest AgriStability Independent Variables Lagged Participation (t-1) AgriStability 0.299*** 0.172*** 2.076*** 0.321*** 0.140*** 2.004*** 0.535** 0.612*** 2.177*** Crop Insurance 2.539*** *** 2.508*** *** 2.826*** ** AgriInvest 0.147*** 1.177*** 0.234*** 0.151** 1.209*** 0.159*** *** 0.374*** Diversification Variables Sector Diversification 0.394* Crop Diversification *** *** *** *** Financial Variables Operating Profit Margin *** ** Operating Expense Ratio 0.057** *** *** 0.104*** *** ** *** Debt Coverage Ratio * *** Previuos Year s Payment *** *** Reliance Previous Year s Business Risk *** *** ** Size Dummies Revenue $ K *** ** 0.121*** 0.159*** Revenue $ K *** ** 0.263*** 0.208*** ** Revenue >$500K 0.116** 0.234*** *** *** 0.270*** Sector Dummies Field Crops 0.222*** ** Others ** ** Supply Managed *** *** Swine 0.326*** *** 0.187* Constant *** *** *** *** *** Rho Crop Insurance-AgriInvest 0.052*** 0.070*** Crop Insurance-AgriStability 0.153*** 0.135*** 0.199*** AgriInvest-AgriStability 0.089*** 0.097*** 0.115* Log-pseudolikelihood -19, , ,459.1 No. of observations 26,138 17,705 2,082 11

13 Table 4. Parameter Estimates from the Multinomial Probit Combination/Independent Variables All Farms Field Crops Beef Combination 1 : Crop Insurance, AgriStability, and AgriInvest Lagged Participation 1.297*** 1.254*** 1.179*** Sector Diversification *** Crop Diversification *** *** ** Operating Profit Margin Operating Expense Ratio * Debt Coverage Ratio *** ** Previous Year s Business Risk *** Previous Year s Drop in Income ** ** Revenue $ K 0.751*** 0.798*** 0.477* Revenue $ K 1.151*** 1.371*** 1.432*** Revenue >$500K 1.778*** 1.391*** 1.086** Constant 1.131*** 1.045*** 1.870*** Combination 2 : Crop Insurance and AgriStability Lagged Participation *** *** *** Sector Diversification ** Crop Diversification *** *** Operating Profit Margin Operating Expense Ratio ** Debt Coverage Ratio *** ** ** Previous Year s Business Risk Previous Year s Drop in Income *** *** Revenue $ K 0.542*** 0.591*** 0.789*** Revenue $ K 0.679*** 0.913*** 1.063** Revenue >$500K 1.335*** 0.826*** 1.622*** Constant 0.743*** 0.781*** 0.880*** Combination 3 : Crop Insurance and AgriInvest Lagged Participation *** *** *** Sector Diversification Crop Diversification *** *** Operating Profit Margin * Operating Expense Ratio *** ** ** Debt Coverage Ratio Previous Year s Business Risk *** *** *** Previous Year s Drop in Income *** Revenue $ K 0.476*** 0.529*** 0.561* Revenue $ K 0.750*** 0.947*** 1.369*** Revenue >$500K 1.247*** 0.819*** 1.037** Constant 0.719*** 1.253*** 2.256*** Combination 4 : AgriStability and AgriInvest Lagged Participation 0.485*** 0.384*** 0.795*** Sector Diversification * Crop Diversification *** *** Operating Profit Margin ** Operating Expense Ratio ** Debt Coverage Ratio Previous Year s Business Risk ** *** 12

14 Previous Year s Drop in Income Revenue $ K 0.391*** 0.396*** Revenue $ K 0.747*** 0.950*** 0.626** Revenue >$500K 1.249*** 0.863*** Constant 1.286*** 0.869*** 0.843* Combination 5 : AgriStability Only Lagged Participation *** *** *** Sector Diversification Crop Diversification *** *** Operating Profit Margin ** Operating Expense Ratio ** Debt Coverage Ratio Previous Year s Business Risk Previous Year s Drop in Income Revenue $ K Revenue $ K 0.313** 0.726*** Revenue >$500K 0.678*** Constant 0.812*** Combination 6 : AgriInvest Only Lagged Participation * *** Sector Diversification Crop Diversification *** *** Operating Profit Margin Operating Expense Ratio *** *** ** Debt Coverage Ratio ** Previous Year s Business Risk *** *** ** Previous Year s Drop in Income Revenue $ K Revenue $ K 0.407*** 0.504** Revenue >$500K 0.670*** 0.369* Constant 0.874*** 1.053*** 0.956* Combination 7 : Crop Insurance Only Lagged Participation *** *** ** Sector Diversification 0.394* 0.550* Crop Diversification *** *** ** Operating Profit Margin Operating Expense Ratio Debt Coverage Ratio Previous Year s Business Risk * ** * Previous Year s Drop in Income Revenue $ K 0.207** 0.197* Revenue $ K 0.440*** 0.629** Revenue >$500K 0.685*** Constant 0.107*** Combination 8 : No Participation (base outcome) Log-pseudolikelihood -34, , ,619.1 No. of observations 26,138 17,705 2,082 13

15 As Table 5 shows, BRM programs are effective at reducing downside risk (results shown for the beef sector) the coefficients for most program participation states are negative and significant, suggesting that producers who participate in one or more of the programs experience smaller drops in gross margin than producers who do participate. One exception is Crop Insurance, which seems to not be able to mitigate downside risk when used independently. Also note that specialization is associated with larger drops in gross margin. Table 5. Parameter Estimates for the Tobit and Quantile Regression Models of the Impact on Program Participation States on Downside Risk Beef Independent Variables Tobit Quantile Regression Q25 Q50 Q75 Participation States Crop Insurance, AgriStability, and AgriInvest * *** *** *** Crop Insurance and AgriStability * *** ** *** Crop Insurance and AgriInvest ** *** *** *** AgriStability and AgriInvest * *** *** *** AgriStability Only * *** *** *** AgriInvest Only ** *** ** *** Crop Insurance Only Diversification Variables Sector Diversification *** 0.204** Crop Diversification *** 0.234** Size Dummies Revenue $ K ** Revenue $ K 1.212** *** Revenue >$500K 1.139* Constant * 2.495*** 3.232*** 8.216*** CONCLUDING DISCUSSION Despite the fact that BRM programs are effective at reducing downside risk, producers exit the programs. Two findings are particularly important in this context: 1) positive and significant state dependence once a producer exits a program, he/she is unlikely to return; and 2) dynamic spillover effects among the programs exit from one program will likely trigger exit from other programs. There are at least two policy implications that can be drawn: 1) it is possible to reduce the risk of low program participation in the future by incentivizing farms to participate in the present short-term policies may have long lasting effects, because they help break the vicious cycle leading to low participation; and 2) policies aimed at incentivizing farmers to participate in one type of program may also contribute in breaking the cycle leading to low participation in the other programs. The question is what are the appropriate incentives to be used? The finding that larger farms more likely to participate across the board suggests that we need to understand what drives participation decisions for the small farms and what can be done to make them sign up. 14

16 REFERENCES Agriculture and Agri-Food Canada ( ), Farm Income, Financial Conditions and Government Assistance Data Book, Heckman, J.J. (1981), The Incidental Parameters Problem and the Problem of Initial Conditions in Estimating a Discrete Time-Discrete Data Stochastic Process, in Structural Analysis of Discrete Data with Econometric Applications, C.F. Manski and D. McFadden (eds.), MIT Press: Cambridge, MA, pp Kimura, S. and J. Antón (2011), Farm income stabilization and risk management: Some lessons from AgriStability program in Canada, Paper presented at the EAAE 2011 Congress Change and Uncertainty: Challenges for Agriculture, Food and Natural Resources, Zurich, Aug 30-Sept 2. Wooldridge, J.M. (2005), Simple Solutions to the Initial Condition Problem in Dynamic, Nonlinear Panel Data Models with Unobserved Heterogeneity, Journal of Applied Econometrics, 20:

17 Appendix 1. Canadian Business Risk Management Programs, Table 1. Review of Canadian Business Risk Management Programs Agricultural Policy Framework ( ) Growing Forward I ( ) Growing Forward II ( ) Crop/Production Insurance AgriInsurance AgriInsurance Canadian Agricultural Income AgriStability AgriStability Stabilization (CAIS) AgriInvest AgriInvest AgriRecovery AgriRecovery Spring Credit Advance Program Advance Payments Program Advance Payments Program Advance Payments Program Agricultural Policy Framework ( ) Canadian Agricultural Income Stabilization (CAIS) meant to help producers protect their farm income from both small and large declines (CAIS integrated income stabilization and disaster protection). Under CAIS, payments were triggered when the program year margin (eligible revenue minus eligible expenses) fell below the historical reference margin (average program margin in three of the past five years, with the highest and the lowest margins dropped). The farmer was not compensated for the entire margin loss; rather, the actual level of payment was based on a series of percentages related to the severity of the margin loss, with deeper losses reimbursed at higher rates, i.e.: - Tier 1: The minimum coverage was 50% if the margin fell in the range from 85% to 100%; - Tier 2: For the next 15% of the margin decline, the government covered 70%, and - Tier 3: The maximum coverage rate was 80% for the disaster portion i.e., margin decline in the range from 70% to 0%. - Also, producers were protected for an amount up to 60% of their negative margin. The producer chose a level of coverage and paid a fee related to coverage level selected (i.e., $4.50 for every $1,000 of reference margin for maximum protection, $3.825 for medium protection, and $3.15 for minimum protection) plus an additional $45 in administration fees. In order to respect World Trade Organization rules, the overall assistance provided by Governments did not exceed 70% of the margin decline at maximum protection (program payout was capped at 66.5% for medium protection and 56% for minimum protection). Production (Crop) Insurance provided crop producers with financial protection against the effects of production losses caused by natural hazards like drought, flood, hail, frost, excessive moisture and diseases. 16

18 Advanced Payment Programs (Spring Credit Advance Program and Advanced Payments Program) provided up to $250,000 in loan guarantees and interest-free loans on up to the first $50,000 advanced to assist in financing the spring seeding and the fall harvest and storage. Growing Forward I ( ) AgriStability replaced the CAIS program, except that tier 1 payments (triggered by the first 15% fall in reference margin) were eliminated and replaced by AgriInvest. Medium and minimum levels of protection were also eliminated (more than 90 percent of producers who had signed up for CAIS chose the maximum coverage level) and program fees were set at $4.50 for every $1,000 of reference margin protected. Producers who experienced negative margins could be compensated for up to 60% of the portion of their margin decline that was below zero. Producers with a negative reference margin were eligible for AgriStability if they had positive margins in two of the three years used to calculate their reference margin. There was no limit on the number of negative margin payments a producer could receive. The maximum payment under AgriStability was capped at either $3 million or 70% of margin decline, whichever was lower. Any amount over this limit would be deducted from the negative margin payment. AgriStability included a number of improvements over CAIS which are summarized in Table 2. Table 2. AgriStability vs. CAIS Program Change Inventory valuation Negative margin coverage expanded Easier participation requirements 2003 CAIS AgriStability Benefit to producers End of year price used to value inventory. Producers with negative reference margins not eligible. Limit on number of negative margin payments a producer could receive. Producers had to make a deposit equal to 22% of their reference margin. Both opening and end of year prices used to value inventory. Producers with negative reference margins eligible. No limit on number of negative margin payments. Producers pay $4.50 per $1,000 of reference margin protected. More accurate assessment of losses. Better protection for those faced with back to back disasters. Cash is not tied up in a deposit. Targeted Targeted advances not Targeted Advance Payment Faster payments in times of 17

19 Program Change advances available Interim payments available available CAIS AgriStability Benefit to producers Interim payments not widely available. in place to provide serious income declines. assistance in times of serious income declines. Interim payments available in most provinces within 30 days. Early access to an estimated portion of payment. AgriInvest is a self-managed producer-government savings account that allows producers to set money aside which can then be used to recover from small income shortfalls (AgriInvest replaces coverage for margin declines of 15% or less, previously covered under CAIS) or finance investments that help mitigate risks or improve market income. Producers can deposit up to 1.5% of their Allowable Net Sales (=Sales of Allowable Commodities (most primary agricultural commodities except those covered under supply management) Purchases of Allowable Commodities) each year into an AgriInvest account and receive a matching government contribution. Producers of supply managed commodities, who also produce allowable commodities, may be eligible for AgriInvest based on the non-supply managed portion of their farming operation. Producers have the flexibility to withdraw funds at any time throughout the year. ANS are limited to a maximum of $1.5 million per eligible participant. The limit on matching government contributions is $22,500 a year. AgriInsurance offers protection for production losses related to specific crops or commodities caused by natural hazards. Premiums for AgriInsurance coverage are cost-shared between the producer, the province and the federal government. Producers receive a payment when their production is below their guaranteed insured level of protection. Compared to Production Insurance, coverage under AgriInsurance includes livestock and additional horticultural crops. However, there is still a need to improve coverage for forage and livestock production. AgriRecovery is a disaster relief framework which allows governments to determine whether or not assistance beyond what is provided through existing programs (AgriStability, AgriInvest, AgriInsurance, etc.) is warranted in response to a disaster. The aim is to help affected producers with the cost of taking actions to mitigate the impacts of the disaster and/or resume business operations as quickly as possible. 18

20 Advance Payments Program provides producers with a cash advance (of max. $400,000 per producer) on the value of their agricultural products during a production period. The federal government pays the interest on the first $100,000 of a cash advance issued to a producer per production period and producers have up until the end of the production period to repay their advance. By improving their cash flow throughout the year, the program provides producers with flexibility for marketing of their commodities (producers can meet their financial obligations and sell when prices are most favourable). The Advance Payments Program combines two previous programs (Spring Credit Advance Program and Advance Payments Program) and includes a number of improvements over those programs: -coverage was expanded to allow loans on additional commodities, including livestock. The APP is delivered under the authority of the Agricultural Marketing Program Act, which was amended in 2007 to make the program available to the livestock sector. -loan and interest free limits were increased to ensure that they reflect existing farm sizes. Growing Forward II ( ) The BRM programs under Growing Forward I have been maintained. However, a number of changes have been made to some of these programs in an attempt to better target them, while placing a higher degree of responsibility on producers to manage and mitigate risk. AgriStability 70% Margin Coverage: Governments will provide assistance once a producer s margin falls below 70% of their historical reference margin (under Growing Forward, the program payout trigger was 85% below the reference margin). This change is meant to target assistance to farm operations facing the most severe income losses. The annual AgriStability fee to protect a farming operation is lowered to reflect this coverage level, at $3.15 for every $1,000 of reference margin. Harmonized Positive and Negative Margin Coverage Levels: A producer s payment will be based on a single level of government support (i.e., 70%) for both positive and negative margin declines. Harmonizing assistance at 70% simplifies the payment calculation and increases assistance in cases of negative margins to those who are eligible. Reference Margin Limit: Payment calculations will be based on the lower of the historical Reference Margin or the average allowable expenses in the years used to calculate the Reference 19

21 Margin. This limit has been introduced to better target AgriStability assistance to more severe income losses, rather than profit fluctuations. AgriInvest Producers can deposit up to 1.0% (instead of 1.5% under Growing Forward ) of their Allowable Net Sales. The limit on matching government contributions will decrease to $15,000 (from $22,500 under Growing Forward) a year. 20