Tax-Exempt Bond Financing for Beginning and Low-Equity Farmers: The Case of 'Aggie Bonds'

Abstract

The U.S. Department of Agriculture (USDA) through the Farm Service Agency (FSA) makes and guarantees loans to beginning farmers who are unable to obtain financing from com- mercial lenders by targeting a portion of its direct and guaranteed farm ownership and operating loan funds to beginning farmers. The program is au- thorized through a provision in the Internal Revenue Code (IRC) covering private activity bonds and relies on private lenders to make loans to eligible farmers; in return, the lender receives a tax exemption on the interest received from the loan.

Headnote

The "Aggie Bond" program was established in the 1980s to provide beginning and low- equity farmers access to capital. The bonds, which pay tax-exempt interest, may be used by qualifying famers for purchases of farm real estate and equipment. Using Aggie Bond data collected from states and Census of Agriculture data spanning 25 years, we examine whether the program has had an impact on farm entrance, land ownership, and the size of operation. We do not find strong evidence that the program led to an increase in the proportion of beginning farmers; however, we find limited evidence the program helped beginning farmers become full land owners as well as increased the rate of growth in the proportion of beginning farmers who are full land owners.

Key Words: Aggie Bonds, beginning farmers, land ownership

JEL Classifications: Q14, Q15, H20, H24

It comes as no surprise that starting a farm op- eration can be an expensive endeavor. Startup requires access to land and capital equipment as well as a commitment of the operator's time. The costs are particularly prohibitive for begin- ning or low-equity farmers. In 2011, the average farm operated 415 acres and held just under $1 million in assets, the vast majority of which was in land and structures.1 On an annual basis, farms may make capital purchases as well. In 2011, 84% of farms made a capital purchase and the average was $36,430." Together, these fig- ures suggest access to capital and its cost present significant hurdles for new entrants.

Ownership of land and capital equipment is not imperative, and it is not uncommon for op- erators to farm a combination of owned and rented land. Furthermore, many of the common capital-intensive tasks of the farm operation can be contracted, for example custom application services for fertilizer, tillage and harvesting, and drying and storing. However, ownership of farmland in particular is attractive because it may be part of a larger wealth-building strategy for the operator. Nickerson et al. (2012) report that farm real estate, which includes structures, accounts for 84% of total value of farm assets. Ownership in farmland represents an investment not only in an asset used for current production, but also as a store of wealth. Operators may wish to own land for other reasons as well. Research has shown that a portion of commodities pay- ments may accrue to landowners, even when they rent their farmland (Kirwan, 2009; Michael, Kirwan, and Hopkins, 2003).

To finance land and capital equipment ac- quisitions, beginning farmers may reach out to traditional lending institutions such as commer- cial banks or farm credit services. However, it is up to them to negotiate terms with the lender, and because they are likely to face constraints as a result of a low equity or cash flow position (which is often the case with new operations), the terms of the loan will reflect such a position.

States and the federal government recognize the need to support agriculture and help new farm operators. The U.S. Department of Agriculture (USDA) through the Farm Service Agency (FSA) makes and guarantees loans to beginning farmers who are unable to obtain financing from com- mercial lenders by targeting a portion of its direct and guaranteed farm ownership and operating loan funds to beginning farmers. FSA targets a portion of its loan funds to small and beginning farmers and ranchers or farmers who have oper- ated a farm or ranch for 10 or fewer years and do not own a farm or ranch greater than 30% of the median size farm in the county.3

A first-time or beginning farmer loan, also known as an 'Aggie Bond," is another source of capital for farmers who wish to establish or ex- pand an operation. The Aggie Bond program was established in the early 1980s to provide eligible farmers and ranchers with loans for purchases of farm real estate and equipment. Initially starting as a pilot program in three states in 1980, Aggie Bond programs currently operate in 16 states (CDFA, 2005). The program is au- thorized through a provision in the Internal Revenue Code (IRC) covering private activity bonds and relies on private lenders to make loans to eligible farmers; in return, the lender receives a tax exemption on the interest received from the loan. Although the program is not costless to federal and state governments-they still lose the revenue associated with the tax exemption on the interest lenders receive-they do not assume the risk of the loan. Instead, the program relies on private lenders to evaluate the creditworthiness of the farmer and to assume the liability of default. The benefit to beginning farmers is that the tax-exempt status of the loan is an incentive to lenders to provide access to credit they might not otherwise provide and at rates that may be below the market rate.

Despite having been around for over 30 years, questions of whether the program has achieved perceived objectives have remained largely unanswered. For example, has the pro- gram encouraged entrance into farming, and does the program affect land ownership or the size of the operation? In this article, we use state Aggie Bond data and Census of Agricul- ture data to examine these questions. We use a difference-in-differences approach to examine the effects of the program on outcome mea- sures, which differ across groups such as U.S. states. Specifically, we evaluate the effects of the Aggie Bond program on beginning farmers in counties for the states where the program was introduced. Using Census of Agriculture data spanning 25 years combined with state Aggie Bond data, we do not find strong evi- dence the program led to an increase in the proportion of beginning farmers; however, we find limited evidence the program may be re- sponsible for a greater proportion of beginning farmers who are full land owners as well as affecting the rate of growth in the proportion of beginning farmers who are full land owners. The next section provides a background on Aggie Bonds and the mechanics of the issuance followed by a presentation of our model and data and finally the results and discussion.

Background

The average age of U.S. farmers and ranchers has been increasing over time. The fastest growing segment of farmers is those 65 years and older. In 1978, 16.4% of principal farm operators were older than 65 years. By 2007, 30% of all farms were operated by producers older than 65 years of age. In comparison, only 8% of self-employed workers in nonagricultural industries in 2007 are that old (Hoppe, MacDonald, and Korb, 2010). Furthermore, farmers aged 55 years and older account for over half of the total value of production, whereas farmers younger than 35 years contribute only 6% of total value of production. This demographic transition has implications for the future of the U.S. agricultural sector.

Ahearn (2011) identified two potential challenges for beginning farmers as having access to land and having access to government programs. As a result, many programs, including government programs and innovative Extension programs, have been introduced to help begin- ning farmers overcome those challenges (Meyer et al., 2011). Although the rest of the literature focuses mostly on management technologies used by beginning farmers (Adhikari, Mishra, and Chintawar, 2009; Mishra, Wilson, and Williams, 2007) or financial performance and financial stress of beginning farmers (Katchova, 2010; Kropp and Katchova, 2011), our study examines a unique tax-exempt bond financing program to help beginning farmers establish their businesses.

Recent evidence from a program similar to the Aggie Bond program offers some potential evidence of the effect of targeted financial as- sistance. Pederson, Chung, and Nel (2012) present findings from a study of Minnesota's Ba- sic Farmer Loan and Seller Assisted Loan (BSAL) programs-two programs that are similar to Aggie Bonds with respect to the qualification restrictions they apply to loan applicants (be- ginning farmer status and net worth limitation, for example). Among their findings, they report Minnesota's BSAL program helped credit con- strained farmers increase their farm's productivity, as measured by gross income, as well as increase their investment in depreciable capital equipment.

Mechanics of Aggie Bond Placement

The tax-exempt status of certain private activity bonds is predicated on the policy goal of address- ing certain failures; a failure in the credit market is one possible explanation for the provisions, while suboptimal provision of a good is an- other. In certain cases, when private business activities are considered to serve the public good (or meet a public purpose test), the proceeds of the bonds may be used for the qualifying purpose and the interest on such bonds is tax- exempt (sections 103 and 141 [e][l] of the IRC). Private activity bonds are authorized for many purposes, including the development of facilities associated with public services such as airports and sewage facilities, manufacturing facilities, certain student loans, redevelopment projects, and small issue farm property.

The Aggie Bond program has been around since a three-state pilot program was established in 1980. Over time, the program expanded to 24 states, although currently only 16 states operate a program. Authority to issue Beginning Farmer Loan Bonds is provided by Sections 141 (Private Activity Bonds) and 144 of the IRC. Under Section 141, certain bonds may be issued and the proceeds used to fund private business activities. The code applies a test to private business, namely more than 10% (5% in some cases) of the proceeds of the issue go to private business use and that the bond is secured by the property used for private activity or payment in respect of the property.

In most cases, Aggie Bond programs are administered by a state development or finance authority, because these agencies are responsible for facilitating other forms of development bond financing.4 For example, Illinois provides bond funding through the Illinois Finance Authority, and Arkansas uses the Arkansas Development Finance Authority. Some states have development authorities specific to agriculture, for example, Iowa.

Unlike other forms of development bonds, the Aggie Bonds are financed by lending in- stitutions and not from an issuance in a public or private bond market. Instead, beginning farmers who wish to acquire Aggie Bond funds start the process by contacting a private lender who par- ticipates in the Aggie Bond program, such as a commercial bank, as if they were pursuing tra- ditional loan financing. Because the lender will hold the bond, the lender evaluates the operator's loan request for risk, collateral, and other conditions, and the lender and operator negotiate terms and conditions of the loan. Once the terms are negotiated, the operator submits an Aggie Bond application with the required material- financial statement information of the operation, background letter, and other information to meet the requirements of the state authority-to the state's Aggie Bond program.

In the bond issuance process, a state au- thority acts as a conduit between the lender and the operator. If the bond application is approved by the authority, the lender purchases the bond from the authority who then lends the proceeds of the bond to the operator. In response, the authority assigns the payment of principal and interest to the lender secured by a promissory note from the operator.

Although the USDA's definition of a first- time or beginning farmer relies on a test of time spent farming, the IRC does not use a measure of time to define such farmers. Instead, it im- poses a test on ownership-either direct or indirect-of farmland. The test limits farmers from owning substantial farmland, which the IRC defines as no more than 30% of the median size of a farm in the county in which such parcel is located (Section 147[c][2][e] of the IRC).

The Aggie Bond program has benefitted from adjustments overtime. First, the Tax Re- form Act of 1986 (TRA86) expanded the term of "beginning farmer" to include farmers who would otherwise qualify except for having pre- viously owned farmland that was disposed of while they were insolvent. TRA86 also allowed bond users to apply 25% ($62,500) of the max- imum bond proceeds to purchase used agricul- tural equipment. Later, the substantial ownership rule was amended in 1996, replacing the 15% of the median rule with the current 30% rule, and bond rules were loosened to allow bond funds to be used to purchase land from related persons (P.L. 140-188, 1996). Before 2008, the IRC placed an additional test on the market value of the land a farmer could own: the market value of the land could not exceed $125,000. In 2008, the nominal maximum bond amount, that is, the largest loan a beginning farmer could access through the Aggie Bond program, was increased. This was arguably the most significant change made to the program since its inception. Starting in June 2008, the maximum bond amount was increased from $250,000 to $450,000 and is indexed for inflation thereafter. Before that, the maximum bond amount had been unchanged; thus, in real terms, the value of the loan has de- clined over time (Figure 1).

Some states impose additional tests for qualifying for a beginning farmer program. In addition to the substantial ownership test for first-time farmers in the IRC, states require ap- plicants to meet a net worth limit test. For ex- ample, Iowa's Beginning Farmer Loan Program requires that farmers have a net worth of no more than $500,000, indexed for inflation by the Pri- ces Paid by Fanners index.5

Econometric Approach

A basic challenge in empirically assessing the impacts of the Aggie Bonds is that we do not observe the counterfactual situation-the out- come of the parties affected by the policy (the treated counties for the states participating in the program) in the absence of the policy. When random assignment is not possible, like in the present case, some method of estimating the counterfactual outcomes such as difference- in-differences is necessary. Another challenge is that we cannot identify in our data which beginning farmers have participated in the Aggie Bond program and which have not. We can only identify the counties (nested in the states) that have implemented Aggie Bond programs. Therefore, our focus is on how the Aggie Bond program has affected the out- comes for all beginning farmers in counties and on states adopting this program rather than on calculating individual-specific effects based on program participation.

Difference-in-Differences Models

Difference-in-differences models typically eval- uate the effects of a program on outcome mea- sures, which differ across groups, such as U.S. states. Specifically, we evaluate the effects of the Aggie Bond Program on beginning farmers in states where the program was introduced. Two differences are commonly considered. One difference is the selection or heterogeneity bias, which is treated by removing the state effects. The second difference is the common economic fluctuations from year to year, which are treated by removing the time effects. The difference- in-differences (DIDs) estimator estimates the effect of the program on the outcome measures after removing the state and time differences. Specifically, two differences are considered: one difference is between the outcomes of in- dividuals located in states participating in the program and those that are not located in states participating in the program, and the second difference in outcomes is between the different periods in terms of the 6 years of census data. The remaining interaction effect (the DIDs) shows the impact of the program on beginning farmers in states that have adopted the program. We aggre- gate the data to county levels to calculate out- come measures affecting beginning farmers.

The DIDs model is defined as

(1) ...

where v" is the outcome measure for county / in one of the six census years t, ti, is a time dummy variable equal to one if the observation is in census year t and zero otherwise, dn is a program state dummy variable equal to one if county / is located in a state participating in the program and zero if it is not, f" a,, is an interaction term be- tween the time dummy variable ti, and active program dummy variable for an active program dit (if the program is active in county / and time t), and xit are other characteristics influencing the outcome variable. The time-active program interaction terms t,, a,, are the DIDs measure for the effect of the program on outcome measures, controlling for common time differences be- tween the two groups. We consider four outcome measures, v" : 1) the proportion of beginning farmers in a county; 2) the proportion of begin- ning farmers who are full land owners in a county; 3) the proportion of beginning farmers who are full land tenants in a county; and 4) the average number of acres operated by beginning farmers in a county. Models are estimated in terms of the levels of these outcome variables as well as growth rates in these outcome variables from the current census period to the next census period. Because the analysis is at the county level, clustered standard errors are calculated with each state serving as a cluster variable with counties clustered within states.

Data

The Census of Agriculture data were used and covered census years 1982, 1987, 1992, 1997, 2002, and 2007. The census identifies farmers based on their years of experience and provides detailed information on their farm businesses. Beginning farmers are defined as those with 10 or less years of farming experience using a USDA definition. Several other commonly used operator and farm business characteristics are used in the analysis. Farmers reported their primary occu- pation as farming if they have spent the majority (50% or more) of their work time on farm or ranch work. Farmers who reported positive number of hours for off-farm work are classified as those who worked off the farm. Farmers who lived on their operation are classified as having farm residence. Government payments are mea- sured as the value of all government payments received. An indicator for a female primary op- erator is also used.

Farm size is measured as operated acres, which is calculated as the acres owned minus the acres rented to others plus the acres rented from others. Full land owners are farmers who own all of the land they operate. Part land owners own some land and rent the rest of the land that they operate. Full land tenants are those who rent all land they operate and do not own land.

Data from the Census of Agriculture is ag- gregated (averaged) at the county level and then aggregated (averaged) again to calculate state level statistics. Because of this intermediate ag- gregation at the county level before calculating state level statistics, the numbers reported in this study use a different calculation method and will differ from the official estimates published by the USDA.

Next, we take a look at the states that had an authorized Aggie Bond program at any time. Table 1 presents the states by year of program activity. In 1982, just four states in our data set had an authorized program: Colorado, Iowa, Illinois, and Nebraska. In 2007, the last avail- able census year, 14 states had such a program. The Aggie Bond data were collected by the authors from the various state agencies that administer the program and were subsequently merged with the census data at the state level. In total, 14 states out of 16 with authorized programs have available data. Table 2 presents descriptive statistics of currently authorized programs. The oldest program in operation is Iowa's Aggie Bond program, and it is also has one of the highest issuance volumes among all program states with more than $4 million of bonds issued per year. On average, the 14 state programs issued just over $2.1 million in loans per year, whereas the average volume ceiling- maximum aggregate amount that could be issued in a year-was $ 12.9 million. In our data, Aggie Bond issuances were never limited by the volume ceiling. The average net worth limit in eight states was $346,592; the other seven states did not have such a limit.

Table 3 provides summary statistics by census year for program and nonprogram states. Pro- gram states are defined as those that have had the Aggie Bond program at least in one of the census years, whereas nonprogram states are defined as those that did not participate in the Aggie Bond program during 1982-2007. In general, in a given census year, program states had a greater pro- portion of beginning farmers than nonprogram states; however, overall both groups show declining proportions of beginning farmers over time. The proportion of beginning farmers who are full land owners is also higher in program states as is the average number of acres operated by beginning farmers.

Results

Two types of DIDs models are used based on both levels and growth in key variables such as proportion of beginning farmers in a county, proportion of full land owners, proportion of full land tenants, and average farm operated acres. We first present the DIDs results for the level of variables (Table 4). The first column of Table 4 presents the results for the outcome measure Proportion of Beginning Farmers in a county. The state program dummy, the variable that indicates whether the county's state had an authorized Aggie Bond program, is negative and robust at the 5% level of significance. The co- efficient tells us that the presence of the program is associated with a 3.4% lower proportion of beginning farmers than in the absence of an Aggie Bond program. The time trend control variables confirm that the proportion of begin- ning farmers is declining overtime. As well, we find the null of no program effect on the pro- portion of beginning farmers in a county of an Aggie Bond state-the DIDs estimators-cannot be rejected for any of the census years.

We find limited evidence that the program may have lead to a greater Proportion of Full Land Ownership and greater Acres Operated (columns ? and D of Table 4). The DIDs esti- mator for census year 2007 indicates that the proportion of beginning farmers who are full land owners is 2.5% higher in counties of states with Aggie Bond programs than in those with- out, and on average, beginning farmers operated 143.3 more acres in the program counties in 2007 as compared with 1982. Both coefficient estimates are robust at the 5% level of statistical significance.

Government payments and the proportion of farmers who also worked off-farm were among other characteristics found to be associated with the outcome measures. All else equal, the larger the amount of average government pay- ments in a county, the smaller the proportion of beginning farmers in a county, which is consistent with the evidence that beginning farmers operate fewer acres eligible for government program payments. Although statistically significant, the effect was small in economic terms-less than 0.002% increase in the proportion of beginning farmers for a $1000 increase in average govern- ment payments that farmers receive in a county On the other hand, a 1% increase in the pro- portion of farmers who also worked off-farm was associated with a 0.119% increase in the pro- portion of beginning farmers in a county

Although we do not find evidence the pro- gram influenced the proportion of beginning farmers in a county, we find limited evidence that the program affects growth rates for the proportion of beginning farmers. One possible explanation for this finding is that we only find a difference in growth rates for one period to the next and perhaps that is not sustained over the years or significant enough to make a difference regarding the proportions (levels). Table 5 presents the DIDs model for growth in variables. Because we calculate growth rates from one year to the next, only five growth rates are calculated from the six census years. The DIDs estimates represent the program effect on the change in growth rates between two pe- riods. For example, the variable Active Pro- gram Dummy 2002 is the effect of the Aggie Bond program on the growth rate between 2002 and 2007 compared with the growth rate of the reference period 1982-1984. In column A, the DIDs estimator for 1997 tells us that the pres- ence of the program decreased the growth of the proportion of beginning farmers by 1.37% between census years 1997 and 2002 compared with the growth rate in the reference period. However, consistent with the change in levels from Table 4, we find evidence of the program's positive effect on the growth rates of the pro- portion of beginning farmers who are full land owners. The DIDs estimator for the change be- tween census years 1987 and 1992 is 1.98% and the change between census years 1997 and 2002 is 2.39%, and both are statistically significant at the 5% level.

Estimates for the growth of the proportion of beginning farmers who are full land tenants show the presence of the program resulted in slower growth over the same census years that found greater growth of full land ownership (Table 5, column C). The presence of the program is associated with a decreased growth rate of the proportion of full land tenants between census years 1987 and 1992 of 1.88%; the change between census years 1997 and 2002 is -2.16%. Both are statistically significant at the 5% level. We cannot reject the null of no program effect in the case of growth of operated acres.

Conclusion

For beginning farmers, starting a farm opera- tion is an expensive endeavor with acquisition of equipment and land as one of the largest hurdles to overcome. The Aggie Bond pro- gram, currently authorized in 16 states, is one lending vehicle beginning farmers may access in coordination of private lenders. In return, the lender receives a tax exemption on the interest received from the loan, providing an incentive to lenders to offer loans to beginning farmers and at rates that may be below the market rate. Although the volume of bonds issued varies significantly across states, it is particularly popular in the Midwest, especially the Corn Belt.

We estimate four models, both for levels and differences in variables, meant to measure the program's impact on the beginning farmer population and aspects of their operation. The results generally show that there is no significant effect of the program on the proportion of be- ginning farmers in a county. We do not find evidence that the program increases the pro- portion of beginning farmers in a county, and in fact, the one significant DIDs estimate showed the proportion of beginning farmers declin- ing in the presence of the program. We find, however, the program has an effect in some census years on the proportion of beginning farmers who are full land owners. In such years, we find the proportion of beginning farmers who are full land owners to be 2.5% greater in the presence of the program. Furthermore, we find the program's presence in a state resulted in the growth of the proportion of full land owners that was 1.98-2.39% greater than growth in a refer- ence period without the program. Although lim- ited, the evidence suggests the program is helping with the acquisition of real property, one of the intended goals of the program.

Those calling for an expansion of the program-primarily by exempting Aggie Bonds from the volume limit applied to private activity bonds6-may wish to consider whether, in light of the evidence we present, the policy objec- tive of encouraging participation by beginning farmers could be achieved in another way. The evidence suggests the Aggie Bond program does not affect the proportion of beginning farmers in a county. On the other hand, the program has a positive effect on full land ownership and acres operated, which addresses the be- ginning farmer's challenges of acquiring land and program acres.

Footnote

1 The term farm means any place from which $1000 or more of agricultural products were produced and sold, or normally would have been sold, during the year.

2 Authors' calculations from 2011 Agricultural Resource Management Survey data from the USDA.

3 Farm size is based on total operated acres as determined by the most current Census for Agricul- ture.

4 The Aggie Bond program in North Dakota is administered by the Bank of North Dakota, which is overseen by the North Dakota Industrial Commission.

5 Internet site: "Prices Paid Surveys and Indexes." http://www.nass.usda.gov/Surveys/Guide_to_NASS_ Surveys/Prices_Paid_and_Prices_Paid_Indexes/index. asp.

6 It should be noted the authors have never found the volume limit to be binding.

References

References

Adhikari, ?., A.K. Mishra, and S. Chintawar. "Adoption of Technology and Its Impact on Profitability of Young and Beginning Farmers: A Quantile Regression Approach." Selected paper. Southern Agricultural Economics As- sociation Annual Meeting, Atlanta, GA, 2009.

Ahearn, M. "Potential Challenges for Beginning Farmers." Choices (New York, N.Y.) 26(2011): 1-6.

CDFA. Aggie Bonds. Council of Development Fi- nance Agencies, 2005. Internet site: www.cdfa.net/ cdfa/cdfaweb.nsf/ordredirect.html?open&id = cdfafactsheetaggiebonds.html. Accessed February 2013.

Hoppe, R.A., J.M. MacDonald, and P. Korb. Small Fauns in the United States: Persistence Under Pressure. EIB-63, U.S. Department of Agricul- ture, Economic Research Service, February 2010.

Katchova, A.L. "An Analysis of the Financial Performance of Beginning Farmers." Selected paper. Agricultural and Applied Economics Association Meeting, Denver, CO, July 25-27, 2010.

Kirwan, B. "The Incidence of U.S. Agricultural Subsidies on Farmland Rental Rates." The Jour- nal of Political Economy 117(2009): 138-64.

Kropp, J., and A.L. Katchova. "The Effect of Direct Payments on Liquidity and Repayment Capacity for Beginning Farmers." Agricultural Finance Review 71(2011):347-65.

Meyer, L., J. Hunter, A.L. Katchova, S. Lovett, D. Thilmany, M. Sullins, and A. Card. "Approach- ing Beginning Farmers as a New Stakeholder for Extension." Choices (New York, N.Y.) 26(2011): 1-7.

Michael, J.R., B. Kirwan, and J. Hopkins. "The Incidence of Government Payments on Agri- cultural Land Rents: The Challenges of Iden- tification." American Joumal of Agricultural Economics 85(2003) :762-69.

Mishra, A.K., C.A. Wilson, and R.P. Williams. "Technology Adoption, Management Practices, and Financial Performance of New and Beginning Farmers: Evidence from a National Survey." Selected paper, American Agricultural Eco- nomics Association Annual Meeting, Portland, OR, 2007.

Nickerson, C., M. Morehart, T. Kuethe, J. Beckman, J. Ifft, and R. Williams. Trends in U.S. Farmland Values and Ownership. EIB-92. U.S. Dept. of Agriculture, Econ. Res. Serv. February 2012.

Pederson, G., W. Chung, and R. Nel. "Micro- Economic Impacts of a State-Funded Farmer Loan Program." Agricultural Finance Review 7(2012): 1-25.

P.L. 140-188. Small Business Job Protection Act of 1996. Public Law (1996).