200353pap

Nick Walraven Federal Reserve BoardWashington DC 20551Nick.A.Walraven@frb.gov

and

Peter BarryUniversity of Illinoisp-barry1@uiuc.eduDraft: September 2003

Abstract

In this paper, we review the prevalence of the use of risk ratings by commercial banks that

participated in the Federal Reserve’s Survey of Terms of Bank Lending to Farmers between 1997and 2002. We find that adoption of risk rating procedures held about steady over the period, witha little less than half the banks on the panel either not using a risk rating system, or reporting thesame rating for all their loans in the survey. However, most of these banks were small, androughly four-fifths of all sample loans carried an informative risk rating. We found that aftercontrolling for the size and performance of the bank and as many nonprice terms of the loan aspossible, banks consistently charged higher rates of interest for the farm loans that they

characterized as riskier, with an average difference in rates between the most risky and least riskyloans of about 1-1/2 percentage points.

The analysis and conclusions in this paper are the authors’ and do not necessarily reflect theopinions of the Board of Governors, the University of Illinois, or their respective staffs. We

thank the participants at the NC-221 Regional Research Committee meeting in Denver Colorado,October 3-4, 2002 for helpful comments and suggestions and Edward Johnson for researchassistance.

Bank Risk Ratings and the Pricing of Agricultural Loans

The management of risk by commercial banks in the U.S. and other developed countrieshas advanced significantly to now address the frequency and severity of loss and an enterprise-wide perspective on credit, market, and operational risks. The goals of risk management are torefine the measures of risk, better match economic capital to the overall risk profile, allocatecapital efficiently among the respective bank enterprises, and to price loans and other productsand services consistent with their marginal contributions to economic capital and risk-adjustedreturns on capital (Matten (2000); Beisses (2002; Smithson (2003); Saunders(1999)). Theproposed New Basel Accord emphasizes these refinements, while offering a menu of choices forinstitutions of different size, operating environments, complexities, and market characteristics. The menu ranges from an expanded set of risk weights compared with those in the 1988 Accordto internal-ratings-based approaches in which institutions estimate probabilities of default and theresulting loss by rating classes of borrowers and loans.

The use of risk-rating systems to summarize multiple features of a bank’s customers orloans has been spreading through the banking system for at least a decade, first among largerbanks, and gradually to medium and smaller banks (Brady, English, and Nelson, 1998; Treacyand Carey, 1998). According to Brady, English, and Nelson, virtually all large banks rated loansin the August 1998 Survey of Terms of Bank Lending by the Federal Reserve System; incontrast, most medium to small banks either did not rate loans or assigned all the loans in thesurvey to a single rating category. Assessing the adoption of these risk, capital, and pricingpractices by banks with different attributes is important to understanding the scope and depth oftheir management resources, their likely adoption of more sophisticated technology in the future,and the design of risk-based capital regulations for safety and soundness in a diverse bankingsystem.

Little contemporary evidence is available about the use of such practices on agriculturalloans (for past studies of credit risk management by agricultural banks, see Barry and Calvert(1983); Moss, Barry and Ellinger (1997); Miller et al (1993); Swackhamer and Doll (1969)). In2003, about half of U.S. commercial banks lending to farmers occurred through smaller

2

community banks that had less than $500 million of assets (ERS,USDA), half came from larger,regional or national banks. The diversity in size and management of banks that lend to farmerssuggests the need for a range of capital management and risk assessment guidelines andcontinued monitoring of structural change in agricultural lending.

This paper reviews the use of risk ratings, risk-adjusted pricing, and other responses tocredit risk by commercial banks when making new agricultural loans. Quarterly data from theFederal Reserve’s Survey of the Terms of Bank Lending to Farmers (STBLF) from August 1997through August 2002 are utilized following the methods employed for business lending byEnglish and Nelson (1998) (hereafter referred to as EN). Bank characteristics, loan pricing, andother risk management tools are summarized, compared to the August 1998 findings of EN, andthen evaluated using regression procedures to determine the effects of risk ratings and other riskcontrol practices on interest rates for new farm loans.

Surveying Banks About Agricultural Lending

The data for our analysis of risk pricing on agricultural loans comes mainly from theFederal Reserve’s Survey of the Terms of Bank Lending to Farmers. This section provides abrief history of that survey and reviews its current scope and the selection of its panel (a moredetailed description of the evolution of the survey may be found in Walraven and Slowinski(1993)).

In 1977, the Federal Reserve Board requested a quarterly survey of banks to gauge thecost, volume, terms, and purpose of credit extended to both commercial businesses and tofarmers. A single longitudinal survey panel of banks was selected to gather information aboutboth types of lending. The survey has been modified since that time, most notably in 1989, whena separate panel of banks was selected to report information on farm loans (some banks remainedon both the business loan panel and the farm loan panel), and in 1998 when questions about theriskiness of loans were added to both surveys.

Since the panel redesign in 1989, a stratified, random sample of 250 insured commercialbanks reports information on each agricultural loan completed during the first week of the second

3

month of each quarter 1. Because the volume of agricultural loans is highly skewed across theuniverse of commercial banks, the first stratum of the survey panel includes ten agriculturallenders that are among the largest holders of agricultural loans. The remaining commercialbanks holding at least $1 million in agricultural loans are divided into four strata, with themembers of each stratum holding successively smaller amounts of farm loans. Sixty banks arechosen randomly from each of these strata.

During the sampling week in each quarter, banks in the survey report the amount, the rateof interest, the maturity, and some non-price terms of each loan that they complete. In recentyears, around 200 sample banks report roughly 4000 loans in each survey.

The prevalence of risk rating in the STBLF panel

About 1/4 of the banks in the panel (48 of 186) for the August 1998 STBLF did not ratethe farm loans that they closed, and almost as many (36 of 186 banks) assigned the same riskrating to all of the survey loans that were reported. Similar to the EN findings, almost all of thebanks in the 1998 survey that either did not assign risk ratings or gave all loans the same riskrating were small banks (less than $1 billion in assets). As a group, these banks accounted forabout 18 percent (739 of 4072) of the total number of farm loans in the August 1998 survey.

Although anecdotes suggest that the use of risk-rating systems has been spreading for alltypes of loans, according to the STBLF panel, the proportion of banks that assigned risk ratingswas little changed during the five years following the EN business loan survey. In the August2002 survey, about 1/5 of the panel (38 of 172 banks) did not rate farm loans, and about 1/4 (42of 172 banks) reported no variation in risk ratings; although in total, these 80 banks closed a littlefewer than 9 percent of the loans reported in the August 2002 survey (440 of 5105 survey loans),a proportion well below the 1998 reading, which gives some support to the assertion that theproportion of farm loans having a risk rating has been growing.

“Agricultural loan” refers to either of the farm loan definitions employed in the

quarterly Report of Condition (Call report). Included are both “loans to finance agriculturalproduction or other loans to farmers” and “loans secured by farm real estate”.

4

1

Most of the banks remain on the survey from one quarter to the next; indeed, 120 of thebanks that reported closing at least one loan in the August 1998 survey also reported a loan inAugust 2002. Among this group of banks, about 50 assigned loans to multiple risk categories inboth 1998 and 2002. This set of banks reported about 2/3 of the number of sample loans (3231of 5105) in the most recent survey. Another 10 banks that reported in both periods did not ratefarm loans in 1998, but had begun to report ratings by 2002. A set of 22 banks that did not ratefarm loans in 1998 still did not rate farm loans in 2002, and another 10 banks had discontinuedrating farm loans by 2002. These 32 banks reported 245 loans in the August 2002 survey. Theremaining 29 banks that reported loans in both 1998 and 2002 assigned all the loans the samerisk rating. Although this singularity is a little unfortunate from an econometric point of view, tosome degree it was inevitable—most banks in this last group reported closing fewer than 5 loansduring the August 2002 sample week, and given that the risk descriptions were designed so thatmost loans fell in the middle of the risk scale, it seems plausible that among only a handful ofloans at a particular bank that they all could be ranked similarly.

Description of risk rating categories

Banks participating in the survey are asked to map their internal risk ratings into a set offive rating categories that are described in detail in the reporting instructions. The loans arecharacterized in terms of the probability of a loss to the bank, rather than the probability of adefault by the borrower. As a result, requirements for compensating balances or collateral canlower the risk rating of an otherwise more risky loan. Loans placed in Category 1, the “minimal”risk category, should bear virtually no chance of loss to the bank. Loans in Category 2, aredescribed as “very unlikely” to result in a loss to the bank. Category 3 loans were termed“moderate risk” and were intended to be an average loan to a typical borrower under averageeconomic conditions. The survey was designed so that most loans would fall in Category 3. Loans placed in Category 4, although still bearing an “acceptable” degree of risk, were in somesense substandard. Category 5 loans were described as “Special mention” loans, such as work-out loans--new loans typically would not fall in this category. Two additional rating categories

5

were provided, the first for banks that rated some loans, but not a particular one that wasreported, while the final designation was for banks that did not rate loans.

Farm loan characteristics by risk rating

August 1998 Survey

In order to compare the EN averages for business loans to the STBLF data, we computedaverages by bank size and risk category that were weighted by the size of the loan and by astratum blowup factor reflecting the ratio of the volume of farm loans outstanding at the panelbank to the volume outstanding at banks not in the survey. As shown in Table 1, panel banks in1998 tended to adjust rates of interest on their loans according to the reported riskiness of theloans in the sense that loans rated the least risky generally had lower rates and those rated mostrisky tended to carry higher rates. However, there was a large degree of variability about thisbroad assertion. Large banks, which are defined as having more than $1 billion in assets, closedtransactions on loans with a risk rating of 3 tended that, on average, carried lower rates of interestthan those with less risky ratings. For medium-sized banks (assets between $1 billion and $100million) and small banks (assets less than $100 million) loans in Category 4 tended to carrylower rates than loans in Category 3. EN found closer correspondence between reportedriskiness of C&I loans and the average interest rate than these averages suggest for farm loans.

6

Table 1

Average Loan Rate by Risk Rating(Weighted by Loan Volume)

August 1998 Survey of Terms of Bank Lending to Farmers

Risk Rating

1

All Large Bank Medium Bank Small Bank

9.32%7.98%9.39%9.33%

29.45%8.83%9.68%9.42%

38.68%7.96%10.22%10.14%

48.93%8.77%9.92%9.86%

59.49%9.10%10.16%10.95%

All9.06%8.44%10.00%9.62%

To the extent that reported rates of interest fail to increase with the reported risk rating,other characteristics of the loan likely compensate the lender for bearing the risk. To examinethis possibility for the August 1998 survey, other reported features of the loans, can be brokenout by risk ratings (Table 2). On average, farm loans in the survey were small; the overallweighted average amount for each loan was $27.3 thousand, with the weighted-average amountincreasing uniformly with the size of the loan from $15.6 thousand for the least risky loans to$79.3 thousand for the most risky loans. In general, loans rated as less risky were more likely tohave collateral associated with them, consistent with the Berger and Udell (1990) hypothesis thatcollateral requirements often offset some of the risk of the loan. Furthermore, less risky loanstended to carry provisions allowing the bank to call the note before maturity, likely affording thebank some protection from post-closing changes in the pattern of interest rates in the generaleconomy. In addition, riskier loans were more likely to have been made under a prior

commitment, which is consistent with Morgan’s (1998) hypothesis that, as economic conditionsworsen (i.e. the general riskiness in the economy increases), lenders make relatively more loansunder preexisting commitments and relatively fewer new loans . Prepayment penalties, althoughrare overall, tended to be more prevalent for loans with risk ratings of 3 or above. Finally, the

7

average maturity of the loans fell with reported riskiness, perhaps suggesting some concernsabout interest rate risk or repayment capability that were not sufficiently assuaged by callprovisions, collateral requirements, and other terms of the loan.

Table 2

Loan Characteristics by Risk Rating

Weighted by Loan Volume

August 1998 Survey of Terms of Bank Lending to Farmers

Risk Rating

1

Amount (thousand $)Percent With CollateralPercent Under CommitmentPercent CallablePercent with PrepaymentPenalty

Average Maturity (months)

21.3

18.6

12.4

5.8

9.4

12.8

18.40.1

24.30.1

14.53.5

5.90.8

9.80.5

14.01.6

15.694.957.0

216.794.870.0

331.261.085.1

453.736.992.1

579.348.094.7

All27.366.880.0

August 2002 Survey

Despite a multitude of changes between 1998 and 2002 among agricultural lenders, theagricultural sector, and the economy as a whole, we examined the August 2002 survey datawithin the same framework as the August 1998 survey. Summary statistics are shown in Table 3. In August of 2002, rates of interest at all sizes of banks showed a more consistent tendency toincrease with the reported riskiness of loans, perhaps reflecting a better use of nonprice terms toadjust the riskiness of the loans than in 1998. For instance, the proportion of loans that were

8

secured rose to more than 90 percent in the August 2002 survey, well above the 67 percent thatwere secured in the survey four years earlier. In addition, loans in the riskier categories weremuch more likely to be secured in the more recent survey. The proportion of survey loans thatthe bank can call prior to the maturity date rose substantially for loans that were of average orlower risk (risk ratings 1 to 3).

Table 3

August 2002 Survey of Terms of Bank Lending to Farmers

(weighted by loan volume)

Risk RatingRates by bank size Large Bank Medium Bank Small Bank All banksLoan Characteristics Amount (thousand $) Percent with Collateral Percent With Commitment Percent Callable Percent with Prepayment Penalty Average Maturity

21.5

11.8

15.9

10.0

5.5

13.2

15.595.076.427.01.8

19.096.273.328.20.4

24.088.674.730.31.1

37.695.193.13.93.5

34.098.892.23.30.5

24.992.780.320.61.8

14.30%5.91%7.04%6.75%

24.40%6.89%7.01%6.86%

34.72%7.19 %7.78 %6.00%

45.11%7.48%8.19%5.39%

56.09%7.63% 9.46%6.54%

All4.99%7.11%7.36%6.05%

Controlling for variations in terms

In this section we use regression analysis to examine the effect of various terms on therate of interest charged by the bank. We include data from all the quarterly surveys from August1998 through May 2002, which provided 84,265 loans. Roughly following EN, we include eitherquantitative or qualitative measures of all the nonprice terms of the loan as explanatory variables

9

for the rate of interest.

The STBLF includes some nonprice indicators that are not included in the STBL survey. First, respondents to the STBLF survey indicate whether the loan is secured by farm real estate,some other type of security, or is unsecured–STBL respondents indicate only whether the loan issecured. Over the entire sample used in this paper, about 8-1/2 percent of loans were secured byfarm real estate, although many of these loans have a maturity that is much shorter than onemight expect for a farm mortgage. The farm survey also asks whether the loan is insured by afederal agency, an important consideration given the various programs available to bankersthrough the Farm Services Agency (formerly the Farmers Home Administration). Finally, thefarm survey asks whether the loan was made in participation with other banks, a traditionalmeans that rural banks use to limit exposure to individual loans.

Our regression specification also differs from EN in that we add bank-specific factors thatmight influence the rate of interest offered on the loan. Banks that maintain a substantialportfolio of agricultural loans differ markedly across various size and performance measures. For example, in the March 2003 Call report2, almost half of farm loans were held by“nonagricultural” banks.3 These institutions typically can diversify risks that they perceive intheir farm loans against those in other parts of their portfolio, perhaps reducing the compensationthey require for more risky loans. We include the ratio of the volume of the bank’s farm loans toits total loans as a right-hand-side variable to control for this type of difference.

In addition, the March Call report also indicates that almost 60 percent of the outstandingvolume of agricultural loans was held by banks with assets of less than $500 million. Thesesmall banks typically depend more heavily than larger banks on depository sources of loanablefunds, and their cost of funding loans likely differs substantially from larger competitors.

2

The Call reports are quarterly statements of financial information that are submitted to

banking regulators. The information reported typically could be found on a bank’s balance sheetor income statement.

In this paper, a nonagricultural bank is one that holds a proportion of agricultural loans

in its loan portfolio that is smaller than the unweighted average of the ratios of agricultural loansto total loans at all commercial banks. In recent years, that average has held around 15 percent oftotal loans.

10

3

Smaller banks also tend to be less diversified geographically, so that they may be morevulnerable than larger banks to adverse local events; for example, drought conditions in only acouple of counties could cause a significant proportion of a small bank’s loan customers to havetrouble meeting scheduled payments, while a larger bank might be less vulnerable to this sort ofshock. To account for these potential differences in our regressions, we add the natural logarithmof the bank’s assets as an explanatory variable.

We include several other ratios that are related to bank performance: 1) return on assets,2) interest expense/ total assets, 3) delinquent loans / total loans, and 4) net charge-offs / totalloans. We also include the ratio of loans to deposits, a traditional indicator of bank liquidity.

Table 4 lists the all of the variables used in the regressions, along with the mean andstandard deviation for each; contrary to the previous tables, these statistics are calculated fromraw, unweighted data. For instance, the mean of the 0-1 indicators shows that 5.4 percent of thesample loans fell in the first (least risky) category, while 41.7 percent were rated in the third(typical risk) category. The average interval until the loan could be repriced was about 3 months,while the average maturity was a little less than one year. Roughly 1/5 of the loans in the samplecould be called by the bank, and very few (about 2 percent) carried a prepayment penalty. Mostof the loans (more than 85 percent) were made under some sort of prior commitment, and morethan 9 out of every 10 loans were secured (although relatively few were secured by farm realestate). The loans tended to come from banks that were more profitable–the average ROA forbanks making the loans was 1.4 percent, a good bit above the 1.1 percent average rate of returnfor all small agricultural banks between 1997 and 2002 (Agricultural Finance Databook). Thedelinquency rate at banks making the loans was about 5-3/4 percent, considerably greater thanthe delinquency rate that prevailed at most agricultural banks.

11

Table 4

Summary of Regression Variables

Variable

Risk Rating 1 (least risk) (1=yes, else 0)Risk Rating 2 (1=yes, else 0)Risk Rating 3 (average risk) (1=yes, else 0) Risk Rating 4 (1=yes, else 0)Risk Rating 5 (most risk) (1=yes, else 0)Risk Rating 7 (bank does not rate farm loans) (1=yes, else 0)Nonprice: Days until loan may be repriced

Nonprice: Days until loan matures (0 if no stated maturity)Nonprice: Call provision (1=yes)Nonprice: Prepayment penalty (1=yes)

Nonprice: Loan made under commitment (1=yes) Nonprice: Loan secured (1=yes)

Nonprice: Loan secured by farm real estate (1=yes)

Nonprice: Loan made in partnership with another bank (1=yes)Nonprice: Loan insured by federal agency (1=yes)Nonprice: Ln (loan amount)Bank: Ln (bank assets)Bank: ROA (percent)

Bank: farm loans/total loans (percent)Bank: interest expense/assets (percent)Bank: all loans/all deposits (percent)Bank: all delinquencies/total loans (percent)Bank: all net charge-offs/total loans (percent)

12

Mean

.054.135.417.224.059.08092.6307..190.020.854.907.085.019.0402.4414.81.4023.42.9985.45.770.35

Std Dev.

.227.341.493.417.235.271327.515..392.141.353.290.278.136.1951.553.18.73024.7.78421.86.670.45

The regression results for the entire sample are shown in Table 5. The adjusted R-squared is 31 percent, indicating that a substantial proportion of the variability in rates of interestreflects differences in the terms of the loans and in bank performance. The T-statistics for mostvariables were significant at the 1 percent level, and the F-statistic of 1616 for the inclusion ofthe full set of explanatory variables also was highly significant.

After controlling for both the nonprice terms of the loan and the bank-specificdifferences, the coefficients for the risk rating indicators suggest a plausible and consistentpricing of loans according to their reported riskiness. For instance, a loan with the least riskyrating, other factors equal, carried a rate of interest that was 1.3 percentage points less than a loanrated the most risky (coefficient on Risk 1 minus coefficient on Risk 5).

Coefficients on most other loan level variables were of a plausible magnitude. Loanswith a prepayment penalty, issued under a prior commitment, issued in participation by morethan one bank, or with federal insurance were priced lower than other loans, consistent with theirrisk-reducing properties. The coefficients on these nonprice terms indicated a reduction of 17 to50 basis points for each characteristic, and all were highly significant.

Secured loans tended to carry a significantly higher rate of interest, which is consistentwith results from Berger and Udell’s (1990) finding that banks tend to extend unsecured loansmainly to their least risky customers. However, the subset of secured loans that were secured byfarm real estate carried substantially lower rates of interest. The reduction in rates associated withreal estate collateral was highly significant, which likely reflects the perception of bankers thatfarm real estate collateral affords considerable insurance against losses on loans.

Among the bank level variables, higher returns on assets were associated with lower rateson farm loans–in other words, more profitable banks in the survey tended to offer lower rates totheir farm borrowers. In addition, banks that specialized in farm lending (as indicated by its farmloan ratio), tended to offer lower rates on their farm loans. While the parameter associated withthe farm loan ratio was highly significant and suggests some economies of scale in making farmloans, the effect was small quantitatively. In contrast, the parameter on the ratio of interestexpense to bank assets, our proxy for the bank’s cost of funds, was large and highly significant.

13

Indeed, the coefficient of about unity suggests that banks tended to pass higher funding costsdirectly to their borrowers. In addition, greater bank liquidity, as measured by the ratio of loansto deposits, was associated with lower rates on new loans.

Among our indicators of portfolio quality, the higher the rate of delinquencies in thebank’s portfolio (total delinquencies, both agricultural loans and other loans), the lower the ratecharged on new loans. Similarly, banks with a higher rate of net charge-offs closed new farmloans with significantly lower rates of interest than other banks in the sample. However, both ofthese indicators are somewhat difficult to interpret because they are backward-looking, whilenew loans necessarily reflect the bank’s assessment of the borrowers prospects in the future.

14

Table 5

Summary of Regression Estimates

Dependent Variable: Effective Rate of Interest Observations: 84265 Adjusted R2: .306 F-statistic: 1616.Variable Intercept

Loan-Level Variables

Risk Rating 1 (least risk) Risk Rating 2 Risk Rating 3 (average risk) Risk Rating 4 Risk Rating 5 (most risk) Risk Rating 7 (not rated by bank) Days until loan may be repriced Days until loan matures Call provision Prepayment penalty Under commitment Secured

Secured by farm real estate In partnership with another bank Insured by federal agency Ln (loan amount)Bank-Level Variables Ln (bank assets) ROA (percent)

Farm loans/total loans (percent) Interest expense/assets (percent) All loans/all deposits (percent) All delinquencies/total loans (percent) All net charge-offs/total loans (percent)

.004-.21-.0071.05-.02-.02-.38

1.01-27.11-18.48111.14-49.34-24.13-29.95

.03 .22 .72.70 1.34.78-.00025.00004.25-.33-.30.13-.44-.50-.17-.11

.75 6.6823.6922.3537.0822.15-12.623.1317.10-9.0-18.447.02-22.63-13.32-6.22-31.04

Parameter

7.55

t-value88.54

15

Of course, economic conditions in the farm sector, as well as those in the broadereconomy, change considerably over time, and it seems reasonable to expect that these changesmight affect the price and terms offered by banks to their farm customers. To explore changes inthe terms of agricultural lending over time, we fit the regression described above separately todata for each quarter. The resulting lengthy list of parameter estimates is given in Table 6. As inthe previous regression, we controlled for as many terms of the loan as possible and for variationin bank characteristics and performance in the sample. As a result, one can examine the changesover time in the spreads for loans of different riskiness. For example, the estimated coefficientfor loans that were rated least risky (risk rating 1) ranged from -0.95 (i.e. 95 basis points) in theMay 1998 survey to 0.04 (and statistically insignificant) in the August 2001 survey. Note thatthese spreads were calculated relative to loans in risk category “6\rated by the bank, although the bank did have a risk-rating program for loans in place (this wasthe risk rating category that was omitted as an explanatory variable in the regressions in order toavoid multicolinearity among the parameter estimates).

16

Table 6

Quarterly Parameter Estimates

Parameter Estimates for the Intercept and Risk Variables

Survey DateInterceptRisk 1Risk 2Risk 3Risk 4Risk 5Risk 7

Aug-9711.9647-0.49308-0.172220.127570.237930.727520.20502

t-stat47.8258-6.53293-2.74252.277313.811618.50172.71945

Nov-9711.783-0.48701-0.230250.030790.222640.694590.08046

t-stat47.8241-4.72214-2.414060.333592.267016.29360.77199

Feb-9812.0506-0.65729-0.47664-0.13870.076270.43728-0.18362

t-stat45.1798-4.73164-3.72419-1.122440.593483.1867-1.44289

May-9812.5247-0.94904-0.59727-0.21326 t-stat44.3514-6.40269-4.21542-1.55301Aug-9812.8159-0.65101-0.39740.10831 t-stat46.4878-5.20316-3.526791.01864Nov-9811.9541-0.49646-0.223820.12527 t-stat41.8561-4.98943-2.609311.60523Feb-9911.4844-0.244090.101690.34907 t-stat42.2632-2.086280.972513.51377May-9911.2163-0.36181-0.14420.31624 t-stat41.6286-3.30496-1.4723.39456Aug-9912.2804-0.64136-0.226020.21224 t-stat39.8524-4.8204-1.87741.84157Nov-9912.3397-0.79417-0.351860.02702 t-stat34.8448-5.92675-2.865140.23316Feb-0013.4864-0.91174-0.72512-0.24191 t-stat39.421-6.66962-5.80215-2.0575May-0011.106-0.51437-0.207670.23163 t-stat42.3539-4.44223-1.949832.31963Aug-0012.3781-0.49469-0.378950.09318 t-stat39.2529-3.66872-3.146930.84679Nov-0011.7332-0.40541-0.068950.35768 t-stat36.2461-3.42432-0.677923.8946Feb-0112.226-0.53378-0.062010.29856 t-stat41.4432-4.7511-0.64273.48386May-0112.0182-0.28291-0.166680.35007 t-stat44.8908-2.98594-2.109035.04267Aug-0111.98020.041190.139990.43473 t-stat36.52860.359061.440844.93891Nov-0110.2524-0.56839-0.232120.19406 t-stat29.1971-4.10386-1.962741.76625Feb-0212.0048-1.11079-0.73761-0.17772 t-stat37.2174-8.55097-6.70699-1.82777May-0213.0292-0.52882-0.55637-0.19007 t-stat46.2952-4.67556-5.84302-2.2415817

-0.14523-1.043250.295882.705810.331934.011680.390193.835830.402954.178670.329652.814120.206881.74272-0.15846-1.320420.335253.304380.262682.413960.486085.217560.431914.964080.636598.902230.683847.742760.388293.531230.183291.884780.10281.202690.326852.19040.653715.42720.657756.53660.829857.21280.687076.58170.950557.65840.563624.38180.277492.08430.563114.92140.597714.68660.888518.25680.863188.46150.712948.15821.0843910.6840.913467.43530.669155.87080.691446.863-0.37056

-2.64859

0.04215

0.37421

0.10106

1.14664

0.34103

3.06162

0.34323

3.28667

0.18731

1.48955

0.10382

0.80211

-0.38769

-2.91743

-0.00855

-0.07642

0.05745

0.45831

0.11482

0.95446

0.06447

0.58647

0.17941

2.1347

0.32605

2.8177

0.37447

2.87293

-0.55219

-4.41506

-0.24112

-2.25384

Table 6 (continued) Parameter Estimates for the Nonprice Terms of the Loan

Survey DateReprice MaturityCallablePrepay Pen.CommitSecurityRE SecureParticipatFed Insur.Amount

Aug-97

-0.00019 t--2.76407Nov-97

-0.00032 t--5.58977Feb-98

-0.0003 t--5.76745May-98

-0.00023 t--4.71837Aug-98

-0.00027 t--4.48298Nov-98

-0.0002 t--3.17592Feb-99

-0.00013 t--2.65477May-99

-0.00013 t--2.51471Aug-99

-0.00028 t--4.91228Nov-99

-0.00026 t--4.00303Feb-00

-0.00039 t--6.47755May-00

-0.00036 t--6.99077Aug-00

-0.00027 t--4.2351Nov-00

-0.00026 t--4.5922Feb-01

-0.00015 t--2.75647

May-012.07E-05

t-0.35862

Aug-010.000114

t-1.97825

Nov-010.000393

t-5.93557

Feb-020.000329

t-5.38693

May-020.000412

t-6.70226

4.75E-060.276290.103977.0784-1.3E-050.26782-0.370316.8464-4E-050.23995-1.211765.71671.52E-050.348950.416459.3999-0.000120.34521-2.957238.5602-2.2E-050.31081-0.568916.957-5.7E-050.1744-1.516243.9498-1.3E-050.11989-0.436053.34061.36E-050.168980.330064.5075-1.6E-050.08725-0.380991.87481.15E-060.419340.032229.24660.0001530.472574.0405612.54752.66E-050.540120.675111.4835-5.3E-050.3997-1.418377.9204-8E-050.16658-2.71063.6348-6.8E-050.05429-2.124921.3039-4E-05-0.10189-1.01662-2.2109-6.5E-05-0.34335-1.56449-6.6449-5.1E-05-0.16862-1.42508-3.22324.17E-05-0.012751.28604

-0.2889

-0.75149-0.090620.05885-7.8939-2.21371.362760.51477-0.320170.064465.2765-7.51961.39694-0.42409-0.359080.01945-1.9762-8.62360.38693-0.40166-0.098220.04622-2.5647-2.31831.000480.58627-0.150880.004014.4534-3.33920.08788-0.10903

-0.16210.0609-0.59-3.40511.26022-0.35993

-0.284890.11349-2.46-6.84382.26891-0.05146-0.229640.19214-0.2896-5.45733.720040.03648-0.35680.049210.199-8.7480.90245-0.40761-0.1990.1336-1.8032-3.94712.12539-0.57311-0.37955-0.0337-2.8126-7.6642-0.49673-0.57029-0.340040.05839-2.8878-7.43031.05269-0.17368-0.331510.06203-0.974-6.23140.96419-1.00168-0.355390.06623-13.6575-6.34741.06715-0.26712-0.36045-0.06803-3.3755-7.2201-1.127310.24669-0.1866-0.189573.3792-3.776-3.989140.57022-0.37818-0.039126.1852-6.8908-0.780931.36634-0.37509-0.0008713.618-6.6978-0.013881.10463-0.35336-0.1209110.4736-6.3012-1.902510.02823-0.61688-0.097680.4013

-12.5623

-1.81107

18

-0.10314-0.11674-1.67555-1.03815-0.15162-0.28636-2.33092-2.93409-0.18019-0.44985-3.09432-4.31259-0.0418-0.56794-0.7631-5.469060.04228-0.159980.65885-1.644570.03771-0.592160.53115-5.17509-0.15161-0.65634-2.38337-6.93978-0.24325-0.4327-4.2119-4.47789-0.14953-0.49377-2.56279-4.498740.07481-0.442831.22758-3.78015-0.08152-0.3555-1.52395-2.90837-0.20064-0.19319-4.43548-1.78079-0.12887-0.1707-2.40827-1.47133-0.19657-0.3162-3.45377-3.10796-0.08393-0.67956-1.62159-6.197070.06327-0.233841.31036-2.66617-0.14094-0.47429-2.64246-4.36047-0.18095-0.60181-3.06528-5.28968-0.0618-0.78589-0.86609-5.83903-0.0678-0.69453-1.10169

-4.92199

0.22815-0.140283.41081-13.61920.02151-0.142560.25447-14.83210.29851-0.121663.46239-12.51960.19577-0.130822.66363-13.34740.14138-0.158891.77816-15.3195-0.14008-0.16425-1.54665-16.21450.15384-0.141681.71461-14.75950.06825-0.127751.04202-13.08560.03082-0.130450.4273-13.00290.31457-0.159083.5399-13.85420.22091-0.111252.44144-9.88580.17802-0.113992.48221-11.09630.15098-0.120481.75042-10.33440.29249-0.159583.2011-15.84450.17289-0.182561.88845-18.82590.01675-0.194680.2347-21.9038-0.0431-0.20218-0.598-20.68280.27328-0.249853.06253-23.8826-0.30933-0.20527-3.84503-19.1277-0.29496-0.20425-4.09403

-21.1152

Table 6 (continued) Parameter Estimates for the Bank-level Variables

Survey Dateln(Bank Assets) ROAFarm RatioInt.Exp./LoansLoan/Dep.Delin. RateNchoff Rate

Aug-97-0.1450.07469-0.0068071 t-stat -11.96131.95324-6.35391Nov-97-0.130080.0559-0.0033028 t-stat -10.55731.25628-2.95617Feb-98-0.124140.05365-0.0013468 t-stat -9.52091.21388-1.25673May-98-0.13807-0.02102-0.000898 t-stat -11.8415-0.68866-0.86343Aug-98-0.15573-0.04038-0.0010973 t-stat -12.1733-1.04452-0.95798Nov-98-0.142950.008085.9908E- t-stat -10.13610.22480.05094Feb-99-0.153040.01385-0.0052002 t-stat -12.41180.40611-4.63371May-99-0.13772-0.01302-0.0054393 t-stat -10.7054-0.42377-5.2472Aug-99-0.182450.01606-0.0024289 t-stat -12.56150.39936-2.2231Nov-99-0.16484-0.06598-0.0013442 t-stat -10.0774-1.29239-1.0454Feb-00-0.16207-0.14525-0.0022265 t-stat -10.3398-2.99953-1.84095May-00-0.082560.10382-0.0032306 t-stat -6.70362.92677-2.98949Aug-00-0.117080.06655-0.0039621 t-stat -7.97331.72145-3.05511Nov-00-0.068330.07666-0.0003842 t-stat -5.1921.79553-0.28297Feb-01-0.146970.14848-0.0036264 t-stat -11.59483.16721-2.96445May-01-0.192060.05987-0.0046814 t-stat -14.8362.97262-3.91116Aug-01-0.20728-0.08752-0.0046762 t-stat -14.186-2.66359-3.35782Nov-01-0.26830.035740.0014206 t-stat -16.89280.96630.9636Feb-02-0.321290.02705-0.0005931 t-stat -20.70530.65255-0.41004May-02-0.32928-0.12912-0.0027003 t-stat -24.8321-3.37411-2.3391219

-0.051450.0055420.0064-1.45344.825571.24193-0.053550.0071620.019063-1.43466.051462.14212-0.046730.0047960.008311-1.26393.517470.83911-0.12970.005327-0.00543-3.67254.35344-3.74318-0.171590.004892-0.020664-4.87623.91853-4.37747-0.03058-0.00027-0.013938-0.8307-0.20487-1.600160.09259-0.001650.0015042.6617-1.315380.163850.09715-0.00128-0.0026852.4993-1.2863-1.90533-0.069160.003577-0.015069-1.77463.143-3.23178-0.029430.002767-0.045947-0.59052.18074-4.34631-0.219550.003949-0.03075-4.82933.32803-2.887830.050620.002073-0.0025281.68582.2436-1.30539-0.061470.0029770.000913-1.67312.642130.131980.02216-0.00081-0.0652590.5578-0.49831-5.742820.0619-0.00132-0.0241021.7661-1.31768-1.733620.04571-0.001160.0082441.0856-1.391133.911980.0836-0.004250.0149031.7608-4.793241.84060.45139-0.0020.03224110.4597-2.679572.614670.28175-0.003870.008117.5936-4.322360.58082-0.09769-0.00051-0.003463-2.3886-0.72799-1.4130.176686.03537

0.080822.03024

0.053861.32545

-0.02559-0.59799

0.113252.88481

0.017510.31657

0.040710.74366

-0.15936-4.05936

0.100731.93752

0.107291.74834

0.23733.23866

0.061862.07688

0.502096.52909

0.507238.54973

0.181053.43665

-0.04869-1.43977

-0.0494-0.93976

0.053730.85016

0.03780.67687

0.040491.00748

It is common in describing financial markets to translate quoted rates of interest into rate spreads

between instruments of varying degrees of riskiness. As shown in Exhibit 1, the markup for loans in either thethird or fourth risk rating category moved closely together, and for the entire sample period averaged about 70basis points over loans in the least risky category. As might be expected, fluctuations in the markup for themost risky loans were much wider than those on other loans, with noticeable spikes in 1999, 2001, and 2002. In general, the spreads largely exhibited similar patterns of ups and downs over the 20 quarters of the timeperiod.

To further examine quarterly changes in the spreads on survey loans of different reported riskiness,

Exhibit 2 compares movements in the estimated spread between survey loans that received the most riskyrating and those with the least risky rating to movements in the spread of speculative-grade issues and thoserated BAA in the corporate credit markets. The spread between the most risky and least risky loans in thisbank survey generally ran a couple of percentage points less than the spread in the corporate markets; althoughthe corporate spreads seem to show a bit of an upward trend over the sample period, while the spread onagricultural loans did not. The quarterly swings in the two spreads seemed to match up pretty well over thesample period, except for a somewhat anomalous spike in the agricultural series in 1999.

20

Exhibit 1

21

Exhibit 2

22

In examining the quarterly parameter estimates, it was apparent that beginning in 2001, the coefficients

on the variables associated with the duration of the loan (the dummy for whether the loan could be called, thedummy for whether the loan could be prepaid, and the number of days until the loan could be repriced), eachswitched sign (the quarterly estimates for these variables are shown in Table 6 and Exhibits 3, 4, and 5 in theappendix). For almost the entire sample period until 2001, the inclusion of a call provision came along withan interest rate that was about 30 basis points higher (significant at the 1 percent level); during this period,about 1/4 of sample loans carried call provisions. Beginning in early 2001, the proportion of sample loanswith call provisions dropped to about 15 percent or less, and the presence of a call provision, on average,reduced the rate on the loan by about 20 basis points. This switch in sign came as short-term interest rateswere falling sharply in the broader economy and economic activity was sagging, perhaps suggesting thatlenders may have been easing terms on their loans to bolster demand. By contrast, the sign and statisticalsignificance of the coefficient for prepayment had not shown any particular pattern until 2001, when theinclusion of a prepayment penalty began to coincide significantly with loan rates that were upwards of a fullpercentage point higher.

In the summer of 2000, the average number of days until the loan could be repriced fell sharply in the

survey. In addition, before that time, a longer period until the loan could be repriced was associated withsignificantly lower rates. However, by the latter part of 2001, fixing the rate for an additional month tended toadd about 10 basis points to the cost of the loan. This effect was highly significant.

Concluding Comments

The data presented here suggest that, among the smaller community banks that provide a substantial

portion of farm loans, the prevalence of risk rating systems changed little during the 5-year (1997-2002)sample period. Nevertheless, there is a sizable volume of farm lending that comes from commercial banks

23

and is carried out utilizing rating systems that enable banks to price the perceived riskiness of their loans. Such risk-adjusted pricing occurs within a framework where loan rates also reflect adjustments for a host ofnon-price terms, including security, commitments, call provisions, prepayment penalties, repricing intervals,and maturity.

The future will likely bring wider use of dual rating systems (frequency of default by borrowers and

severity of default associated with loan transactions), as well as closer linkages between loan pricing, creditrisk, economic capital, and risk-adjusted returns on capital. The systematic pricing practices on farm loansfound in this study thus provide a benchmark to future research on loan pricing as the structure and managingof banks’ credit risks continues to evolve. The resulting spreads between loans that are rated to be ofminimal risk and those of high risk change over time in a pattern that is broadly consistent with qualityspreads in corporate credit markets. Thus, the risk and other pricing characteristics of farm loans largelyparallel those of non-farm business loans.

24

References

Agricultural Finance Databook, Statistical Release E.15, Federal Reserve Board, various issues.www.federalreserve.gov/releases

Barry, P.J. and J.D. Calvert. “Loan Pricing and Profitability Analysis by Agricultural Banks.” AgriculturalFinance Review 43 (1983): 21-29.Brady, T.F., W.B. English and W.R. Nelson. “Recent Changes to the Federal Reserve’s Survey of Terms ofBank Lending.” Federal Research Bulletin 84 (August 1998): 604-615.Berger, Allen, and Greg Udell (1990), “Collateral, Loan Quality, and Bank Risk”, Journal of MonetaryEconomics.Bessis, J. Risk Management in Banking 2nd ed., New York: John Wiley & Sons 2002.Brady, T.F., W.B. English and W.R. Nelson. “Recent Changes to the Federal Reserve’s Survey of Terms ofBank Lending.” Federal Reserve Bulletin 84 (August 1998): 604-615.English, William, and William Nelson (1998), “Bank Risk Rating of Business Loans”, Finance andEconomics Discussion Series, 1998-51, Federal Reserve BoardMatten, C. Managing Bank Capital: Capital Allocation and Performance Measurement, 2nd ed., New York:John Wiley & Sons, 2000.

Miller, L.H., P.N. Ellinger, P.J. Barry and K. Lajili. “Price and Non Price Management of Agricultural CreditRisk.” Agricultural Finance Review 53 (1993): 28-41.Moss, L.M., P.J.Barry, and P.N.Ellinger \"The Competitive Environment for Agricultural Bankers in the U.S.\" Agribusiness: An International Journal 13(July/August 1997):431-444

Morgan, Donald (1998), “The Credit Effects of Monetary Policy: Evidence Using Loan Commitments,” Journal of Money, Credit, and BankingSaunders, A. and Lina Allen. Credit Risk Measurement, 2nd ed., New York: John Wiley & Sons, 1999.Smithson, Charles. Credit Portfolio Management, New York: John Wiley & Sons, 2003.Swackhamer, G. and R. Doll. Financing Modern Agriculture: Banking Problems and Challenges, FederalReserve Bank of Kansas City, 1969.

Treacy, William, and Mark Carey (1998), “Credit Risk Rating at Large U.S. Banks”, Federal ReserveBulletin, November 1998Walraven, Nick, and Sam Slowinski (1993), “Surveying Banks About Agricultural Lending”, Proceedings ofthe International Conference on Establishment Surveys, June 27-30, 1993. American Statistical Association.

25

Exhibit A-1

26

Exhibit A-2

27

Exhibit A-3

28