Progress 10/01/03 to 09/30/09
Outputs
OUTPUTS: This project focused on Farms Service Agency (FSA) direct and guaranteed loans. Earlier in the project, guaranteed loans behavior by banks was examined to identify those factors that motivated commercial banks to utilize FSA guaranteed loans as well as the interest assistance programs that can be used on FSA guaranteed operating loans. Results indicate that larger banks with higher loan to asset and agricultural loan to total loan ratios were more likely to use FSA loan guarantees. FSA also provides an interest assistance program whereby borrowers who would not otherwise be able to cash flow their business plans are given a four percentage point reduction in interest rates. A double hurdle model was estimated to predict which banks making guaranteed operating loans would provide the interest assistance option on a guaranteed loan as well as modeling how many interest assistance loans a bank would make. Results indicate bank characteristics, farm and non-farm financial characteristics, regional location, and time were relevant variables in explaining bank usage of the interest assistance program and intensity of use. A 2004 study supported by FSA generated large data sets about the direct loan program. The study initiating the data collection analyzed the effectiveness of the FSA direct loan programs. The study addressed issues of the adequacy of serving the potential market for direct loans and if direct loans were going to the intended groups, graduation progress of direct loan borrowers and measures of the FSA direct loan subsidy rates and how the rate could be lowered. The general study findings are that FSA ha good market penetration during the study years in that loans were serving financially stressed, family farms. Findings also show that a majority of direct loan borrowers in fiscal years 1994-1996 had exited the direct loan program by November 2004. Most of the terminating borrowers continued framing using either commercial credit or no credit. Some borrowers graduated to FSA guaranteed loans. More recently interest focused on using Cox regression methods to estimate hazard functions for direct, FSA, operating loan times to loan termination. Empirically many loans were observed performing beyond their initial maturity dates without having registered a monetary loss. This happens in large part because loans can be restructured and have up to fifteen years to pay back under the restructuring terms. The empirical modeling captured this behavior by allowing loans to terminate in one of three ways: default, payment-in-full or restructured. The restructured loans were subsequently modeled as separate sample to predict the time to termination where loans could terminate via default or payment-in-full, or considered as censored if they had not terminated. Analysis in the models focused on four independent variables: loan-to-value, loan amount, debt-to-asset ratio, and repayment ability. Results indicated that all four variables were significant in at least some estimated models. Time-varying covariates representing contemporary economic conditions were also significant. PARTICIPANTS: Dixon, Bruce L. Professor, University of Arkansas, Fayetteville, Department of Agricultural Economics and Agribusiness, Division of Agriculture. Ahrendsen, Bruce L. Professor, University of Arkansas, Fayetteville, Department of Agricultural Economics and Agribusiness, Division of Agriculture. Danforth, Diana M. Program Associate. University of Arkansas, Fayetteville, Department of Agricultural Economics and Agribusiness, Division of Agriculture. Martini, S. Program Associate. University of Arkansas, Fayetteville, Department of Agricultural Economics and Agribusiness, Division of Agriculture. Chavez, Eddie. Program Associate. University of Arkansas, Fayetteville, Department of Agricultural Economics and Agribusiness, Division of Agriculture. McFadden, Brandon. Graduate Assistant. University of Arkansas, Fayetteville, Department of Agricultural Economics and Agribusiness, Division of Agriculture. Landerito, Aiko. Graduate Assistant. University of Arkansas, Fayetteville, Department of Agricultural Economics and Agribusiness, Division of Agriculture. TARGET AUDIENCES: The primary target audiences are the academic community and government agencies. The primary academic community is composed of those researchers in agricultural finance. Publications and presentations were made in journals read by these scholars and meetings where these scholars convene. USDA's Farm Service Agency administers the direct and guaranteed loan programs to farm operators and was the primary target audience of the research. PROJECT MODIFICATIONS: Not relevant to this project.
Impacts
The outcomes were presentations, theses, refereed publications and a report. The project outcomes were findings of statistical significance from various studies described above. Many of the study results were presented at meetings of the NC1014 and NC1177, Agricultural and Rural Finance Markets in Transition. These are annual meetings attended primarily by academic professionals and some government researchers involved in agricultural credit studies and provisions. An overriding goal of the Farm Service Agency (FSA) direct and guaranteed loans programs is to fill credit market gaps to make credit available to agricultural borrowers on the financial edge and to support lending institutions and local communities. The studies developed under this project contributed to the continuing dialogue among other academics and government researchers and administrators studying these programs. As noted in the Outputs section, this project contributed substantially to the goal of investigating the FSA direct loan program effectiveness. A number of articles and presentations extended the work done under the Effectiveness study which was funded directly by FSA. The main Effectiveness study report was cited in a report to Congress by Charles Dodson and Stephen Koenig (Evaluating the Relative Cost Effectiveness of the Farm Service Agency's Farm Loan Programs, August 2006). In addition that report cited a study by one of the principals who worked in conjunction with an Economics Research Service agricultural economist. That study examined issues in farm bankruptcies and farmer exits over the twentieth century. The study provides data and analysis key to this continuing issue about the structure of agriculture. So the research efforts of the present project have definite policy relevance and were brought to the attention of the national legislative body. Research using the data bases developed in the Effectiveness study has further developed results relevant to program effectiveness. Resources' provided by the project allowed the principal investigators to allocated their time to the various studies undertaken. In particular this allowed the principals to advise four masters theses associated with the project and one Ph.D. dissertation.
Publications
- Chavez, Eddie C., Bruce L. Dixon, Bruce L. Ahrendsen, and Eric J. Wailes. 2009. Comparative Financial Characteristics of U.S. Farms by Type, 2005. AgEcon Research in Agricultural & Applied Economics. Staff Paper SP 02 2009, University of Arkansas Department of Agricultural Economics and Agribusiness, Fayetteville, Arkansas. Accessible at: http://purl.umn.edu/55780.
- Landerito, A., Dixon, B. L., B. L. Ahrendsen, D. Danforth and S. Hamm Analyzing FSA Direct Loan Borrower Payback Histories: Predictors of Financial Improvement and Loan Servicing Actions. Selected paper presented at AAEA Annual Meetings, Milwaukee WI. July 26-28, 2009. Accessible at: http://ageconsearch.umn.edu/handle/49340
- McFadden, Brandon, Dixon, B. L., B. L. Ahrendsen, D. Danforth, and S. H. Martini. Impact of Time to Default and Post Origination Events on FSA Direct Loan Losses. Paper presented at the NC-1177 Agricultural and Rural Finance Markets in Transition. Chicago IL, October 1-2, 2009. Proceedings in Press.
- McFadden, Brandon. Impact of Time to Default and contemporaneous Events on FSA Direct Loans Losses. Unpublished MS thesis, University of Arkansas, Fayetteville, December 2009.
- Landerito, Aiko. Determinants of Financial Improvement and Loan Servicing Actions for FSA Direct Loan Borrowers. Unpublished MS thesis, University of Arkansas, Fayetteville, August 2009.
|
Progress 01/01/08 to 12/31/08
Outputs
OUTPUTS: The research associated with this project was also part of the research activities of the Center for Farm and Rural Business Finance at the University of Arkansas and research projects ARK02025 and ARK01745. Research focused primarily on estimation of hazard functions for USDA Farm Service Agency (FSA) direct loans originated in fiscal years 1994-1996. The Farm Service Agency makes direct loans to farm operators on the financial margin. The major purpose of direct loans is to provide credit to financially stressed borrowers who are nonetheless creditworthy. Hazard functions give the probability of some event given that the object of interest has survived to a given point in time. Hazard functions were estimated in a competing risks framework using Cox regressions. In this case the two possible outcomes are default or paid in full. In the first part of the study, hazard functions were estimated as dependent only on data available at the time of loan origination. In a later phase of the study variables that changed over the duration of a loan were introduced into the model. These were variables describing the general economic environment of the state or region of the borrower. The pre-origination data were collected by FSA for a larger study of the effectiveness of the FSA direct loan program. Results of the first part of the study were presented at NC-1014and published in proceedings in January 2008. The second part of the study is currently under review. In addition, two articles were disseminated in the Farm Management and Marketing Newsletter published by the University of Arkansas Cooperative Extension Service. The first article was a macroeconomic outlook for 2008 and included topics such as national income, unemployment, farmland prices and input costs. A second article addressed the likelihood of another agricultural financial crisis given the then meteoric rise of agricultural commodity prices. Both of these papers were directed at policy makers, farm operators and agricultural business executives. PARTICIPANTS: Bruce L. Dixon and Bruce L. Ahrendsen, Professors, Diana Danforth, and Sandy Hamm, Program Associates, Brandon McFadden, Graduate Assistanct and Monica Foianini, former Graduate Assistant, Department of Agricultural Economics & Agribusiness, University of Arkansas, Division of Agriculture, Fayetteville AR; Steven Koenig and Charles Dodson of the Farm Service Agency also provided substantial background information and assistance in obtaining data for the analyses. TARGET AUDIENCES: The findings of this research have useful information for lenders to farm operators, landlords, farm managers, policy makers, industry, governmental program managers and other research and extension professionals PROJECT MODIFICATIONS: Essentially no changes.
Impacts
University of Arkansas researchers worked with a survey group and professionals at the Farm Service Agency. The team collected data on a sample of direct loan borrowers who closed direct one-year and seven-year operating loans during fiscal years 1994-1996. These data were all borrower and loan characteristics as of loan origination date. This data set was supplemented with another data set of time-varying variables including characteristics of the farm economy and farm operators observed at the regional or state level. A major hypothesis was that events post-origination should have an impact on the outcome of a loan. The time of origination variables were unique to the borrower and loan. The time-varying characteristics described the economic environment on a quarterly basis and allowed for the impact of unknowable future economic events during the life of the loan. Many of the standard, pre-origination variables had an impact as expected. The borrower's debt-to-asset ratio increasing was associated with shorter times to default and longer times to be paid in full, as expected. Relative default risk rose dramatically for both loan types as loan amount increased. This finding is surprising since loans should be set at an amount the borrower can be expected to pay. The study also revealed that the strength of past association with FSA was important. Borrowers who had more outstanding FSA direct loans at time of origination were less likely to default for both loan maturities. The explanation for this is that FSA possesses knowledge of the borrower that is not reflected in the observable data collected and so is able to make better evaluations of borrower creditworthiness. A surprising finding was that the time-varying covariates on farming and economic trends (observed quarterly) had strong impacts on loans with one-year maturities but only one time-varying covariate was significant for loans with seven-year maturities. We had expected that seven-year loans would be more affected by post-loan origination events than one-year loans. One explanation is that in the sample the one-year loans were for larger amounts than the seven year loans. This makes periodic debt service for seven-year loans much lower than for one-year loans implying a short term event may be more influential for shorter loan maturities. The analysis here points in at least two directions. The findings indicate that post-origination events are important in determining a loan outcome. First, in evaluating a loan's viability pre-origination, an analysis of loan hazards could be made to determine the sensitivity of a loan to default as a function of projected economic events. Second, strength of prior relationship with FSA is significant in predicting default behavior. The reasons for this relationship should be further investigated.
Publications
- Dixon, B. L., B. L. Ahrendsen, M. Foianini, S. Hamm, and D. Danforth. 2008. Competing Risk Proportional Hazard Models of Farm Service Agency Direct Operating Loans. In Agricultural and Rural Finance Markets in Transition. ed. Michael A. Gunderson, University of Florida, Department of Food and Resource Economics, Proceedings of Regional Research Committee NC-1014 (January):208-230. Dixon, B. L., and B. L. Ahrendsen. 2008. Is Another Agricultural Financial Crisis Looming Farm Management and Marketing Newsletter. University of Arkansas, Division of Agriculture. 16,2(June):2-4. Dixon, B. L., B. L. Ahrendsen, and M. A. Foianini. 2008. Macroeconomic Impacts on Agriculture. Farm Management and Marketing Newsletter. University of Arkansas, Division of Agriculture. 16,1(March):3-6.
|
Progress 01/01/07 to 12/31/07
Outputs
OUTPUTS: Outputs contained reports of analysis on Farm Service Agency direct loans. One refereed journal article was generated and two papers were presented. The paper presentations were at the Hawaii International Conference on Business and the NC-1014 Agricultural Markets in Transition meetings in St. Louis. A masters thesis was defended and is now in final preparation for deposition in the University of Arkansas library.
PARTICIPANTS: Bruce L. Ahrendsen, Professor, Department of Agricultural Economics & Agribusiness, University of Arkansas, Division of Agriculture, Fayetteville AR. Diana M. Danforth, Program Associate, Department of Agricultural Economics & Agribusiness, University of Arkansas, Division of Agriculture, Fayetteville AR. Sandy Hamm, Program Associate, Department of Agricultural Economics & Agribusiness, University of Arkansas, Division of Agriculture, Fayetteville AR. Monica Foianini, Graduate Assistant, Department of Agricultural Economics & Agribusiness, University of Arkansas, Division of Agriculture, Fayetteville AR. Steven Koenig and Charles Dodson of the Farm Service Agency also provided substantial background information and assistance in obtaining the data for the analyses.
TARGET AUDIENCES: Other scientists in agicultural finance.
Impacts
Data collected in 2004 under another project in collaboration with the Farm Service Agency (FSA) were vital in the two major projects pursued this year. These data were refined into a data set that was used for both a multinomial logit model and duration analysis. Both projects examined various aspects of Farm Service Agency direct loans. In particular, FSA is concerned with borrowers becoming lifetime borrowers. The policy goal of FSA direct loans is to provide credit where credit gaps arise in agricultural credit markets. A further policy is for such credit to be transitional and serve as a conduit to conventional credit markets such as the FSA guaranteed loan program. An important part of the project was to determine if the direct loan program is successful in helping farmers to graduate to conventional credit sources. The project was also concerned with identifying whether borrowers continued in production agriculture. A multinomial logit model was estimated with
four possible event types: borrower was still active in FSA direct loan programs, borrower was still operating a farm using conventional credit sources or no credit at all, borrower voluntarily left farming, or that the borrower left farming involuntarily (other than death). The variables influencing which of these outcomes was the most likely were demographic variables, type of loan (operating loan, farm ownership loan or emergency loan) and whether loan was in the socially disadvantaged or beginning farmer program, variables giving the number of existing direct loans when the observed loan was closed, and various financial characteristics. Results also showed that most borrowers were not becoming permanent FSA clients and which can be considered a degree of program success. The second project, which is still in progress, uses duration (time-to-failure) models to identify those variables that have a significant influence on whether a loan defaulted or was paid in full and the time it
took to get to this outcome. This involved the use of what are called competing risks models in the duration literature. Results are preliminary but indicate that financial strength is important in indicating type of outcome and time to that outcome. So is the degree to which the borrower was connected to FSA via the number of direct loans outstanding when the loan under study was originated. This project is the subject of a graduate student's master thesis. The project supported both principals in their involvement on the project.
Publications
- Dixon, B. L., B. L. Ahrendsen, M. Foianini, S. Hamm and D. Danforth. Competing Risk Duration Models of Farm Service Agency Direct Loans. In Hawaii International Conference on Business 2007 Conference Proceedings, http://www.hicbusiness.org (Presented in Honolulu HI, May 24-27, 2007):776-778.
- Dixon, B . L.. B. L. Ahrendsen, O. J. Nwoha, S. J. Hamm, D. D. Danforth. FSA Direct Farm Loan Program Graduation Rates and Reasons for Exiting. Journal of Agricultural and Applied Economics. 39,3(December 2007):471-488.
- Dixon, B. L., B. L. Ahrendsen, M. Foianini, S. Hamm and D. Danforth. Competing Risk Proportional Hazard Models of Farm Service Agency Direct Operating Loans. in Agricultural and Rural Finance Markets in Transition. Presented at Regional Research Committee NC-1014, St. Louis M., October 4-5, 2007.
|
Progress 01/01/06 to 12/31/06
Outputs
The work associated with this project was also part of the research agenda of the Center for Farm and Rural Business Finance at the University of Arkansas and a regional research project and was reported in the research reports for projects ARK0145, ARK02024, and ARK02025. Work continued on Farm Service Agency (FSA) loan program research with two projects focused on guaranteed loans. FSA guaranteed loan program is the major credit source sponsored by the government to production agriculture. The government guarantees up to 95 percent of principal and interest to commercial lenders in the event of default. The first study estimated a double hurdle econometric model of commercial bank use of loan guarantees. The first equation of the double hurdle model indicates whether the bank made a guaranteed loan in the given fiscal year. The second equation predicts how much the bank lent given it made a loan. Models were estimated for three loan categories: all loans, loans to
socially disadvantaged borrowers and loans to beginning farmers. The data were annual, panel in nature and included almost all banks in the U.S. Lenders with larger asset size, higher loan-to-asset and agricultural loan-to-total ratios, and a multi-bank holding company affiliation were more likely to use the FSA guaranteed loan program to serve their farm borrowers. Results also indicated that past major users of guaranteed loans also made a higher volume of principal in the present. But this major lender status did not necessarily indicate that banks would lend larger volumes to beginning and socially disadvantaged farmers. In the second study the impact of the interest assistance program on commercial lender use of guaranteed loans was estimated. Interest assistance is a program whereby the lender can reduce the interest rate on a guaranteed loan by four percentage points and be reimbursed by the government. National panel data were used from the first study. A variation of the
double hurdle model was estimated to predict which banks among those making guaranteed loans would also use interest assistance. As in the double hurdle model, a probit equation was estimated. The second equation was a count data model estimating how many interest assistance loans a bank would make given that it made at least one loan. The results indicate bank characteristics, farm and non-farm financial characteristics, regional location, and year are important factors in determining bank usage of the interest assistance program and its intensity. But many differences remain unexplained by the estimated models. Both studies lack variables to reflect regional demand variations, particularly with respect to number of farmers which varies dramatically across states. Activities are currently underway to obtain Census of Agriculture data to estimate the number of eligible farm operators from state to state. These data will be inserted into the models of both studies. It is expected such
data will increase model explanatory power and perhaps make currently insignificant variables significant due to a better specified model.
Impacts
Guaranteed loans are likely to continue as the primary means of government provision of credit to production agriculture. It is always a government goal to provide such programs at low cost. By focusing on bank characteristics such as asset size, it might be possible to encourage large banks to become more active making guaranteed loans and to become preferred lenders, thereby making guaranteed loans at lower cost. Reasons for large regional variability in interest assistance use bears further investigation.
Publications
- Ahrendsen, Bruce L., O. John Nwoha, Bruce L. Dixon, Daniel M. Settlage, and Eddie C. Chavez. FSA Direct Loan Targeting: Successful and Financially Necessary? Agricultural Finance Review 67,1(Spring 2007):in press.
- Ahrendsen, Bruce L., Charles B. Dodson, Bruce L. Dixon, and Steven R. Koenig. Research on USDA Farm Credit Programs: Past, Present, and Future. Agricultural Finance Review 65,2(Fall 2005):165-181.
- Ahrendsen, Bruce L., Bruce L. Dixon, O. John Nwoha, Sandra J. Hamm, and Diana D. Danforth. Analysis of Farm Service Agency Direct Loan Loss Likelihoods and Loss Rates. AgEcon Search, Research in Agricultural and Applied Economics, agecon.lib.umn.edu (2006):29 pp.
- Ahrendsen, Bruce L., Bruce L. Dixon, Chris Bacchus, and Latisha A. Settlage. Commercial Bank Usage of the Farm Service Agency Interest Assistance Program. In Agricultural and Rural Finance Markets in Transition. ed. Ani L. Katchova, University of Illinois at Urbana-Champaign, Department of Agricultural and Consumer Economics, www.ace.uiuc.edu/agfin/proceedings/AgFin2006.pdf (December 2006):40-61.
- Dixon, Bruce L., and Bruce L. Ahrendsen. Macroeconomic Impacts on Agriculture. Farm Management and Marketing Newsletter. University of Arkansas, Division of Agriculture 14,1(March 2006):1-4.
- Dixon, Bruce L., Bruce L. Ahrendsen, O. John Nwoha, Sandra J. Hamm, and Diana D. Danforth. FSA Direct Farm Loan Program Graduation Rates and Reasons for Exiting. AgEcon Search, Research in Agricultural and Applied Economics, agecon.lib.umn.edu (2006):30 pp.
- Nwoha, John, Bruce L. Ahrendsen, Bruce L. Dixon, Eddie C. Chavez, and Daniel M. Settlage. Farm Service Agency Direct Farm Loan Volumes and Market Penetration by Farm Size, Socially disadvantaged and Beginning Farmer Cohorts, 2000-2003. In Agricultural and Rural Finance Markets in Transition. ed. Christine Wilson, Purdue University, Department of Agricultural Economics, Staff Paper #06-03, www.ace.uiuc.edu/agfin/proceedings/AgFin2005.pdf (March 2006):11-40.
- Settlage, Latisha A., Bruce L. Dixon, Bruce L. Ahrendsen, and Steve R. Koenig. Factors Determining the Use of Guaranteed Loans by U.S. Commercial Banks. AgEcon Search, Research in Agricultural and Applied Economics, published at: agecon.lib.umn.edu (2006):30 pp.
- Settlage, Latisha, Bruce Dixon, Bruce Ahrendsen, and Steve Koenig. A Detailed Look at Lender Participation in the Farm Service Agency Guaranteed Loan Program. In Agricultural and Rural Finance Markets in Transition. ed. Ani L. Katchova, University of Illinois at Urbana-Champaign, Department of Agricultural and Consumer Economics, www.ace.uiuc.edu/agfin/proceedings/AgFin2006.pdf (December 2006):23-39.
|
Progress 01/01/05 to 12/31/05
Outputs
Activities were advanced in three areas: landlord selection of land-leasing contract type and terms; graduation rates and ways of Farm Service Agency (FSA) direct loan borrowers from direct farm loan programs to other sources of credit; the propensity of commercial banks to originate FSA guaranteed loans. The landlord lease selection study was completed with the publication of the final results about selection of contract type and terms. This study utilized data from a survey of 706 Arkansas landlords that yielded 199 usable questionnaires. Econometric models were estimated predicting which type of contract (rent, crop share or cost share) that landlords would select. Three different models were estimated according to three different theoretical paradigms. These paradigms were credit constraints, agency problems and a risk aversion model. The data rejected the agency model but the evidence between the credit constraint and the risk aversion models was not as clear.
The models indicate landlord risk preferences play an important role in lease type selection. Also, the models indicate that credit constraints exert some influence on lease type selection. Results also indicate that crop share contract terms are functions of land and crop characteristics. The study of FSA direct loan graduation rates utilized a survey of 2715 questionnaires from a nationwide survey of FSA direct loan borrowers who had initiated loans during fiscal years 1994-1996. FSA farm loan managers reported the method of exit from programs utilized by past borrowers who were no longer FSA direct loan participants. Borrowers could graduate to guaranteed loans, regular commercial loans or no loans at all and still be farming. They also could have exited to leave farming voluntarily or involuntarily. A multinomial logit model was estimated to identify those variables statistically significant in selecting the type of exit or if the borrower was still in the farm loan program.
Borrowers with strong financial backgrounds (lower debt-to-asset ratios, higher net worth and more non-farm income relative to gross cash receipts) were more likely to graduate from FSA programs. This was expected but also indicates that lowering eligibility requirements could lead to longer stays in direct loan programs and more voluntary and involuntary exits from farming. For the study of commercial bank use of guaranteed loans, a portfolio selection model from the theoretical literature on banking was modified to include an asset choice for guaranteed loans. A massive data base was constructed for the purposes of this study. Data were obtained from FSA identifying all the guaranteed loans made by commercial banks from 1995-2003. The involved banks were identified by an appropriate code so that data on the bank could be collected from the FDIC Call Report data for that bank. Data were also collected from the Summary of Deposits data source. Bank asset size, loan-to-asset ratio,
agricultural loan-to-total loan ratio, and multi-bank holding affiliation were among the lender variables considered as explanatory variables in the regression models. Estimation results are still being analyzed.
Impacts
The study of landlord lease selection makes clearer what types of factors motivate landlord selection of lease types and terms. This greater understanding of landlord behavior should facilitate better negotiations between landlords and tenants and demonstrates the importance of studying both landlords and tenants in lease studies. Research also indicates that Farm Service Agency borrowers do leave federal farm loan programs and that stronger initial eligibility requirements lead to higher graduation rates. But such practices might negate part of the intent of the direct loan program.
Publications
- Settlage, Latisha A. Examining the Use of Farm Service Agency Guaranteed Loans by Commercial Banks. Unpublished Ph.D. dissertation; Purdue University , August 2005.
- Rainey, R. L., B. L. Dixon, B. L. Ahrendsen, L. D. Parsch, R. W. Bierlen. Arkansas Landlord Selection of Land-Leasing Contract Type and Terms. International Food and Agribusiness Management Review. 8:1(2005):1-19.
- Nwoha, J. B. L. Ahrendsen, B. L. Dixon, E. C. Chavez, S. J. Hamm, D. M. Settlage, and D. Danforth. Farm Service Agency Direct Farm Loan Program Effectiveness Study. Research Report #977, Arkansas Agricultural Experiment Station, Division of Agriculture, University of Arkansas System, Fayetteville AR, December 2005.
|
Progress 01/01/04 to 12/30/04
Outputs
Activities on Chapter 12 bankruptcy came to completion. In one phase an econometric model of state-level Chapter 12 bankruptcy filing rates was published. A fixed effects model using annual data on 48 states from 1987-2000 was estimated. Independent variables included measures of farm financial characteristics, structural measures of farm size and farm enterprises, government payments and variables representing social forces to keep commitments. The study provides a complement to previous bankruptcy studies of reorganization filings (prior studies had been completed by other researchers for bankruptcy Chapters 11 and 13). Results indicated that financial factors and a pent-up demand in 1987 were the major predictors of Chapter 12 filing rate fluctuations. The study emphasizes the careful interpretation of the impact of government payments. While these payments were negatively associated with filing rates, it must be remembered that both borrowers and creditors
anticipate payments so that removal of payments might also lessen lending activity and therefore default. It might be the unanticipated payments that lead to the negative association with filing rates. In a related study the history of farm bankruptcy and farm exits was analyzed. This study tabulates the farm bankruptcies from 1898 through 2002. It is noted that the introduction of the Bankruptcy Code in 1979 led to the demise of a complete set of farm bankruptcy statistics. The only reliable farm bankruptcy data post-1979 are the Chapter 12 filings which are an incomplete listing since farmers can also file in Chapters 7, 11 and 13. The large surges during the twentieth century in farm bankruptcies were during the 1920s and Great Depression and then the farm financial crisis of the early 1980s. While farm numbers declined from a high in 1935 of 6.8 million to 1.9 million in 1997, farmers leaving via bankruptcy are a minority of the exits. Farm numbers did not even necessarily decline
when farm bankruptcies were high as in the 1920s. In fact the major exodus from farming began post World War II when farming was relatively prosperous. A unique aspect of the research is a special tabulation from the Administrative Office of the U.S. Courts on the discharge, conversion and dismissal rates of Chapter 12 filers. These data show that Chapter 12 files do not necessarily receive discharges. In fact, the rate of discharge is declining over time. That is, cases filed in the early years of Chapter 12 were more likely to receive discharges that those filed in the 1990s. This can be attributed to creditors and debtors becoming more familiar with Chapter 12 and wanting to settle outside the formal confines of Chapter 12 to minimize costs on both sides. A final activity was the analysis of inferential accuracy of statistical event studies using daily futures returns for corn, soybeans, live cattle and hogs. Regression models were compared with constant mean return models. At
small levels of abnormal performance the GARCH(1,1) model with a t distribution was the consistently more powerful model of those tested.
Impacts
While farm bankruptcies were numerous over two periods in the twentieth century, the twenties and thirties and then the eighties, they were not the primary reasons for declining farm numbers over the twentieth century. Farm bankruptcies appear to lag behind the movement of farm prices, farm income and other economic conditions that cause bankruptcy. Farm bankruptcy accounted, proportionately, for a small number of farm exits. Institutional changes in the eighties and nineties in the form of loan mediation, borrower rights, and risk management have given farmers additional means of dealing with financial stress.
Publications
- Dixon, Bruce L., Angela Ma, Bruce L. Ahrendsen, Latisha Settlage, Jerome M. Stam. Factors Affecting State-Level Chapter 12 Filing Rates: A Panel Data Model. Emory Bankruptcy Developments Journal. 20:2(Spring 2004):401-426.
- McKenzie, Andrew M., Michael R. Thomsen, Bruce L. Dixon. The Performance of Event Study Approaches Using Daily Commodity Futures Returns. Journal of Futures Markets. 24(2004):533-555.
- Stam, Jerome M. and Bruce L. Dixon. Farmer Bankruptcies and Farm Exits in the Untied States, 1899-2002. USDA/ERS Agriculture Information Bulletin No. 788, Washington D.C., March 2004.
|
Progress 01/01/03 to 12/31/03
Outputs
An investigation has been completed that identified the factors explaining the differential filing rates of Chapter 12 across states in the U.S. A regression model was estimated using fixed effects on panel data. The data were for the lower 48 states using annual data from 1987-2000. Previous studies had shown that both economic and social norm factors were important in explaining filing rates for Chapter 7 and Chapter 12 bankruptcies. The present study shows that financial factors are more important in understanding variation in Chapter 12 filing rates over time and across states than social norms. Additional Chapter 12 research was conducted on farm bankruptcies and exits during the twentieth century. Comparisons of Chapter 12 filing rates from 1986-2002 were made with overall farm bankruptcy rates from 1898-1979,the pre-Chapter 12 era. The evidence suggests that farmers were more likely to file for bankruptcy in the nineties than in earlier decades of comparative
economic well-being. Farmer bankruptcy is not the typical way farmers exit farming. Data on the disposition of Chapter 12 cases were also examined. Chapter 12 cases filed in 1986 and 1987 were more likely to receive discharges than cases filed in later years. This suggests that Chapter 12 is providing a framework for debtors and creditors to resolve their differences in out-of-court settlements to the benefit of both parties. It may be that a higher proportion of "hard" cases are being filed in Chapter 12. As a lower proportion of cases were discharged, a corresponding increase in the proportion of cases dismissed arose. The proportion of cases converted to other bankruptcy chapters remained fairly constant over time. In a land leasing study of landlords leasing land in Arkansas, research on landlord participation and views of the 1996 FAIR Act were conducted. One study identified the determinants of landlord satisfaction with lease terms. Satisfaction was measured on an ordinal scale
and ordinally, ranked, limited dependent variable models were estimated to measure the impact of various variables on landlord lease satisfaction. The sampled landlords leased cropland from the five crop reporting districts that are the primary crop producing regions of Arkansas. Results of that study indicated that proportion of income from leasing, presence of irrigation and tenant education were important determinants of lessor satisfaction. Landlord perceptions that the FAIR Act favored landlords was positively associated with landlord satisfaction. A study with data on both the landlord and the tenant for the same parcel would be a rich data source.
Impacts
Study shows that the farmer bankruptcy chapter, Chapter 12, is likely forming a framework used by debtors and creditors for out-of-court debt settlements. Thus passage of current bankruptcy legislation to make Chapter 12 a permanent part of the Bankruptcy code would have a long-term impact on farm debtor-creditor negotiations.
Publications
- Dixon, B. L., Ma Nan, B. L. Ahrendsen, Latisha Settlage, Jerome M. Stam. "Factors Affecting State-Level Chapter 12 Filing Rates: A Panel Data Model." Bankruptcy Developments Journal. In Press. Rainey, R. L. , B. L. Dixon, L. D. Parsch, B. L. Ahrendsen and R. Bierlen. "Landlord Satisfaction with Arkansas Agricultural Land Agreements." Journal of Agricultural and Applied Economics. 35(2003):543-553. Stam, J. M., B. L. Dixon and W. Rule. "Sixteen Years of Chapter 12 Bankruptcy: Evolution of Filing and Disposition Rates." Agricultural Finance Review. 63(2003):93-108.
|
Progress 01/01/02 to 12/31/02
Outputs
Chapter 12 bankruptcy was enacted in 1986. It has always been a temporary part of the Bankruptcy Code. It was proposed to become a part of the reform bankruptcy legislation that failed in the 107th Congress. Although Chapter 12 filing rates have declined from their record setting levels in 1987 and 1988, filing rates are still high by historical standards as are all forms of bankruptcy filings during the nineties to the present day. The research investigates what causes variation in Chapter 12 filing rates across states and years. Data were collected on Chapter 12 filing rates from 1986 through 2001 at the state level. The filing rate per ten thousand farms was computed. Regression models were estimated to explain the variation over time and states of these rates. It was hypothesized that there are essentially four categories of variables that explain variation in rates: (1) economic variables such as debt-to-asset ratio, net farm income, unemployment rates etc., (2)
variables describing the farm structure such as farm size and proportion of a state's farm income from crops, (3) government policy represented by government payments to farmers, and (4) social variables such as proportion of population divorced and proportion of farm operators over the age of 65. Preliminary findings indicate several variables are important in explaining the rate of Chapter 12 filing. Debt-to-asset ratios are important. As expected, as debt load relative to asset values increase, debtors are faced with increasing financial stress and possible default. This variable also can reflect falling asset values so that the relief from Chapter 12 might be large since secured debts can be written down to current market value of assets under the provisions of Chapter 12. Off-farm sources of income are also important. Both number of days worked off farm and the state level unemployment rate are important and significant. What this likely means is that farm operators are
supporting many unprofitable operations and are willing to work off farm to support the enterprise. Considerable research in the past has been undertaken about tenant participation in land leases and what motivates their behavior. Similar information is lacking for landlords. AASS generated a partial list of landlords in the process of collecting information for the 1997 agricultural census. From the list, 706 landlords who likely had leases in five Arkansas crop-reporting districts were surveyed. One hundred twenty-five usable questionnaires were returned. Ordered probit models were estimated to determine which variables significantly impact satisfaction levels. Estimated models show that the type of lease (cash, share or cropshare) is not a significant indicator of landlord level of satisfaction. Proportion of landlord's income from leasing, tenant's education, social closeness with tenant, presence of irrigation equipment and beliefs about the FAIR Act prove to be significant
variables for lease satisfaction in one or more of the three satisfaction models estimated.
Impacts
Congress has been wrestling with a bankruptcy reform act for the last five years. Currently Chapter 12 is legislated to sunset on June 30, 2003. Our results show Chapter 12 is used in times of financial stress and that government payments are important to farmers in reacting to financial stress. Our results indicate that different factors affect landlords and farm operators in terms of satisfaction. Studies of landlord-tenant relations should include both landlords and tenants.
Publications
- Ahrendsen, Bruce L., Eric J. Wailes, Bruce L. Dixon, Michael Popp. Pat Manning and Tony E. Windham. Arkansas Agriculture 2002 Situation and Outlook. Arkansas Agricultural Experiment Station, Division of Agriculture, University of Arkansas, Research Series 493, June 2002.
- Ahrendsen, Bruce L. And Bruce L. Dixon. "Possible Dealings with Debt." Farm Management and Marketing Newsletter. Arkansas Cooperative Extension Service, University of Arkansas. 10(September 2002):5-6.
- Mutondo. J. E. Lucas D. Parsch, Bruce L. Dixon, Bruce L. Ahrendsen. "Survey Results on the Characteristics of Rice Leases in Arkansas."Farm Management and Marketing Newsletter. Arkansas Cooperative Extension Service, University of Arkansas. 10(September 2002):6-8.
- Dixon, Bruce L and Bruce L. Ahrendsen. "Macroeconomic Impacts on Agriculture." Farm Management and Marketing Newsletter. Arkansas Cooperative Extension Service, University of Arkansas. 10(March 2002):2-3.
|
Progress 01/01/01 to 12/31/01
Outputs
Leased land is a vital component of production agriculture in the United States. It is not only a means by which far operators obtain land to operate, it is also a method of financing production agriculture and sharing the risk inherent in agricultural production. The empirical study of leasing relations is analyzed infrequently due to the problems of gathering suitable data. In existing empirical studies the subjects are typically the farm operators. In the present study landlords leasing parcels in central and eastern Arkansas were surveyed about their largest lease. The survey resulted in just over 300 usable survey instruments. Analysis of the data and model estimation are ongoing. Ordered probit and standard regression models were hypothesized and estimated to identify those factors most important in determining the type of contract selected and the terms of those contracts. The terms predicted by the regression models are the cash rent level, the percentages of
crop shares and the percentages of cost shares. The hypotheses of credit constraints, agency problem and risk aversion are used in specifying alternative models of contract type selection. Results suggest credit constraints are a viable paradigm for predicting contract type selected and the results certainly do not rule out risk aversion. In addition, managerial ability and social capital are also supported as important factors in explaining contract type choice. In the terms' models, land and crop characteristics are important factors in predicting the size of crop terms. Variables for selection bias are not significant. In investigating the satisfaction of landlords with their contracts, 127 observations were judged usable. In the regression models estimated the dependent variable is ordinally ranked to indicate the level of lease satisfaction. Ordinally ranked, limited dependent variable models are used to test for the significance of variables hypothesized to affect satisfaction
levels. Regression results show that type of lease (cash rent, crop share or cost share) is not an important factor in determining lease satisfaction. Further regression results indicate that proportion of landlord's income from leasing, tenant educational background, social capital, existence of irrigation equipment and perceptions about the Federal Agriculture Improvement and Reform (FAIR) Act of 1996 impact lease satisfaction. A comparison with an earlier sample of Arkansas farm operators shows that landlords are more satisfied with their leases than farm operators.
Impacts
42. Impact The passage of the FAIR Act of 1996 changed the support payment mechanisms considerably for production agriculture. Previously deficiency payments and target prices were tied to specific crops and agreements to grow certain types of crops. Under FAIR, payments are no longer tied to specific crops produced and payments are instead tied to specific parcels of land and paid to those bearing production risk. Consequently landlords and tenants are in a changed negotiating framework. Our results indicate that a large majority of landlords did not believe that FAIR changed the distribution of lease income between operators and landlords. Selection of lease type appears unaffected by perceptions about the FAIR Act. Our results indicate that different factors affect landlords and farm operators in terms of satisfaction and lease type selection. This suggests that studies of landlord-tenant relations should include both landlords and tenants.
Publications
- Ahrendsen, Bruce L. and Bruce L. Dixon. "Arkansas Agricultural Interest Rates for 2001." Farm Management and Marketing Newsletter. Arkansas Cooperative Extension Service, University of Arkansas. 9(March 2001):5-6.
- Bierlen, R., L. N. Langemeier, B. L. Dixon and B. L. Ahrendsen. "Land Leasing and Debt on Farms: Substitutes or Complements?" Quarterly Journal of Business and Economics. 39(2000):18-38.
- Rainey, R. L., B. L. Dixon, L. D. Parsch, B. L. Ahrendsen, R. Bierlen. "Landlord Satisfaction With Arkansas Agricultural Land Leases." Selected paper at SAEA annual meetings, Ft. Worth TX, January 28-31, 2001.
- Rainey, R. L., B. L. Dixon, L. D. Parsch, B. L. Ahrendsen, R. Bierlen. "Arkansas Landlord Selection of Land-Leasing Contract Type and Terms." Selected paper at AAEA annual meetings, Chicago IL, August 5-8, 2001.
- Dixon, Bruce L and Bruce L. Ahrendsen. "Macroeconomic Impacts on Agriculture." Farm Management and Marketing Newsletter. Arkansas Cooperative Extension Service, University of Arkansas. 9(March 2001):4-5.
|
Progress 01/01/00 to 12/31/00
Outputs
The 1996 Federal Agriculture Improvement and Reform (FAIR) Act decisively changed the way the federal government supports production agriculture. Under the act, the seven crops benefiting from deficiency payments under the old regime no longer receive these payments. Instead, production flexibility contract (PFC) payments are paid to those individuals bearing some or all of the production risk. Hence landlords cash-renting their land to a farm operator would receive none of the PFC payment but could try to extract part of the payment by increasing rents. Additionally, FAIR also removed acreage restrictions for most crops so that farmers could grow those crops dictated by profit and agronomic motives more freely than in the past. A survey of landlords leasing land in eastern Arkansas and the river valley was conducted at the end of 1998 and the beginning of 1999 to determine how leasing practices had changed since 1995, the last year under the old policy. Landlords
were asked general information about their degree of involvement in negotiations and their opinion of the FAIR Act, and they were asked to describe the terms of their largest cotton, soybean, or rice lease. Data were tabulated and regression models estimated to explain the changing nature of leasing practices. Five regression models were estimated to test hypotheses about land leasing arrangements and the effect of the 1996 FAIR Act on their arrangement. Binomial and ordered probit models were estimated for the five leasing impacts examined in this study: contract choice, cropping mix, lease satisfaction, lease adjustments, and FAIR impacts. One of the unique features of the study is that the respondents were landlords as opposed to tenants. In past survey work on land leasing arrangements, the respondents traditionally have been tenants. In this study the respondents were landlords and were divided into two categories. The first category included what are called pure landlords. These
landlords operated 80 or fewer acres. The maintained hypothesis was that these lessors had less information about agriculture than lessors who were actively involved in agriculture. Thus the other group of landlords were those landlords operating 80 or more acres of land. The results show that the two groups of landlords had distinctly different responses and factors that controlled their responses in the five regression models. That is, the regression independent variables and the estimated coefficients were distinctly different for the two sub-samples of landlords. Three hundred ten usable questionnaires were returned with about one-fourth from operator landlords and three-fourths from pure landlords.
Impacts
Results showed that the major impact of the FAIR Act on leasing is the change in the crop mix on leased land. On at least 31% of the leases, there was some change in crop mix. Only 8% of the leases had some adjustment in either rent levels or share percentages. In regard to the overall impact of the FAIR Act as to whether the Act favored landlords or tenants unequally, the landlords largely believed that neither party was favored over the other. Of the 310 sampled landlords, 84.5% believed that there was no impact or had no opinion. The remaining landlords were almost evenly split, with about 8% saying landlords were favored or tenants were favored. These results are important because they indicate no strong feelings about the FAIR Act in terms of one particular group being favored over another. An earlier national study based on a relatively small and non-random group of producers and professional farm managers came to different conclusions. When Congress meets in
2002 to fashion new farm legislation, our study indicates that landlord versus tenant equity should not be a major issue. In general the results for pure landlords provided better information, probably due to the larger sample size.
Publications
- Settlage, Daniel, B. L. Dixon, and M. R. Thomsen. "A Comparison of Various Frontier Estimation Approaches Under Differing Data Generating Processes."Journal of Agricultural and Applied Economics. 32(2000):410. (Abstract).
- Settlage, Latisha, B. L. Dixon, and B. L. Ahrendsen. "Factors Determining FSA Guarantee Loan Loss Claim Rates in the U.S. for 1990-1997." Journal of Agricultural and Applied Economics. 32(2000):393. (Abstract)
- Bierlen, R. L. D. Parsch and B. L. Dixon. "Testing Cropland Contract Decision-Making Hypotheses: Evidence from an Arkansas Tenant Survey." International Food and Agribusiness Management Review. 2(1)(1999):103-121.
- Bierlen, Ralph, Lucas D. Parsch, Bruce L. Dixon and Bruce L. Ahrendsen. "The1996 FAIR Act: Measuring the Impacts on Land Leasing." Review of Agricultural Economics 22(2000):336-354.
- Bierlen, R., L. N. Langemeier, B. L. Ahrendsen, and B. L. Dixon a "Land Leasing and Debt on Farms: Substitutes or Complements?" Quarterly Journal of Business and Economics. 39:2(2000):18-38.
- Toyne, M. P., J. A. Millar and B. L. Dixon. "The Effect of CEO Control on the Risk of CEO Compensation." Journal of Corporate Finance. 6(2000):291-306.
|
Progress 01/01/99 to 12/31/99
Outputs
The major source of Farm Service Agency (FSA) credit to production agriculture is the guaranteed loan program. In this program Farm Service Agency (FSA) guarantees up to 95 percent of loan principal to private sector lenders. However, these loss payments are a cost to the U.S. Treasury. Data were collected to model loss claims rates for the FSA guaranteed loan program. A statistical model was constructed using data from 40 of the major farm lending states in the U.S. over the time period 1990-1997. Regression modeling techniques were utilized to identify those factors and characteristics of the farm, macroeconomic and agribusiness environment that most influenced the rate of loss claims paid to the actual outstanding principal. Two models were estimated: one for operating loans and a different model for farm ownership loans which are for longer term debt. Data were provided by several FSA offices to form a unique data set that did not previously exist in any readily
obtainable form. Annual data from 1990-1997 were used at the state level. Forty states comprised the sample resulting in a panel data set. Both models were estimated assuming fixed effects by state. Because over 35 percent of the observations on the dependent variable in the farm ownership model had values of zero, the models was estimated by a Tobit estimator. Preliminary results indicate that farmer financial condition is very important in predicating loss claims. Moreover, structure of agriculture appears to be important as well. Interest rate level is also important as is the presence of banks specializing in agricultural loans. Further analysis is continuing. A different project consider investigated the relative accuracy of different methods of frontier function estimation. The research question was to determine the comparative accuracy of two methods of estimation: (1) maximum likelihood estimation, and (2) corrected ordinary least squares. In addition, the comparative accuracy
of three popular approximating forms: (1) the translog, (2) generalized Leontief, and (3) Cobb-Douglas was examined. These questions were answered for a variety of data generating processes that included different distributions of production inefficiency, operator behavioral assumptions and curvature of technology. Results indicate that the simplest approximating form, the Cobb-Douglas was superior. Choice of estimator was not as important although maximum likelihood was better. Producers who optimize knowing their level of inefficiency are difficult to model.
Impacts
The analysis confirms that FSA loss claims should increase with declining farm income. Results also indicate some re-configuration of government support for agriculture might help those operators using guaranteed loans. In particular, it might be useful to consider ways in which the interest rate assistance program could be more effective.
Publications
- Settlage, Daniel M. "A Comparison of Various Stochastic Frontiers Under Differing Data Generation Assumptions." Unpublished M.S. thesis, University of Arkansas, Fayetteville, 1999,
- Fultz, Latisha A. "Factors Determining FSA Guaranteed Loan Loss Claim Activity in the U.S. for 1990-1997." Unpublished M.S. thesis, University of Arkansas, Fayetteville, 1999.
- Dixon, Bruce L., Bruce L. Ahrendsen and Scott M. McCollum. "Models of FSA Guaranteed Loan Use Volume and Loss Claims Among Arkansas Commercial Banks." Arkansas Agricultural Experiment Station Research Bulletin 962. Division of Agriculture, University of Arkansas, November 1999.
- Dixon, Bruce L. Damon McKelvey, Travis Rogers, Frank L. Farmer and Daniel M. Settlage. "Farm Operator Satisfaction with Retail Pesticide Suppliers in the Arkansas Delta." Arkansas Agricultural Experiment Station Research Bulletin 960. Division of Agriculture, University of Arkansas, February 1999.
|
Progress 01/01/98 to 12/31/98
Outputs
Two areas were investigated: (1) modeling of Farm Service Agency(FSA) loss claims activity on guaranteed loans on a nationwide basis and (2) analysis of bankruptcy activity in the mid 1990s. To date the modeling of FSA loss claims activity on a nationwide basis has involved a large data collection effort. We are gathering annual data from 1989-1998 on a state-by-state basis for the lower forty-eight states from a variety of sources. FSA offices in Kansas City and Washington, D.C. have been most cooperative in providing data on loan obligation activities, debt outstanding, volume of delinquent loans, interest assistance and loss claims levels. When compiled, the data set will be able to provide a detailed analysis of guaranteed loan and loss claims activity. Also, data are being obtained from various public sources on a variety of variables that indicate farm operator income both on the farm and off-farm, variables describing the general state of the economy and
variables such as number of agricultural bankruptcies, rate of return on farm assets and rate of return on equity. The purpose of the modeling is to identify state level financial variables that explain the variation in the proportion of loss claims paid during a year to principal outstanding a the the beginning of that year. The model will help policy makers better understand the workings of the guaranteed loan program and how loss claims can be better managed. A phenomenon gaining considerable public attention in the mid 1990s is the large number of individual bankruptcies. Analysis was undertaken to understand the causes of the increases in individual bankruptcies. The conclusion was that at a national level, the easy availability of credit cards might be a major explanatory factor since the typical, past suspected causes like unemployment or downturns in the economy are not correlated with the rates of individual bankruptcy. Further analysis of bankruptcy data shows that Chapter
12 filings did not increase over the mid 1990s although personal bankruptcies did. An important question is whether farmers are filing in record numbers along with the U.S. population overall. Since farmer bankruptcies are not recorded as farm bankruptcies unless filed in Chapter 12, the question cannot be definitively resolved. However, it is known that in the past farmers have utilized Chapter 7 (liquidation) and Chapter 13 even when Chapter 12 was available. So it is likely that they are using these forms today. More study is needed to determine if consumer debt, as opposed to farm debt, is leading farmers into bankruptcy.
Impacts
(N/A)
Publications
- Dixon, B. L. and D. M. Settlage. Consumer Bankruptcy Filings in the U.S. and Arkansas: Growth Rates and Possible Causes. 1998. Arkansas Law Notes. 1998: 13-20.
|
|