Impact of the 1996 FAIR Act on Major Agricultural Input Suppliers

Jian Yang and David J. Leatham

Jian Yang is a graduate student and David J. Leatham is a professor, both in the Department of Agricultural Economics, Texas A&M University. We would like to thank John Bizjak, Donald R. Fraser, David A. Dubofsky, C. Richard Shumway, Edward G. Smith, and two anonymous reviewers for their help and comments. This study reports research conducted through the Texas Agricultural Experiment Station, the Texas A&M University System.

Abstract

A multivariate regression model is used to evaluate the effects of events leading to the passage of the Federal Agricultural Improvement and Reform (FAIR) Act on the returns to shareholders of two major agricultural input industries. The evidence suggests that passage of the act has produced significant positive abnormal returns to the shareholders of large agricultural chemical firms. Conversely, shareholders of small agricultural chemical firms, and large or small farm machinery firms have not experienced significant abnormal returns.

Key words: FAIR Act, event study, input suppliers.

Article <top>

On April 4, 1996, the Federal Agricultural Improvement and Reform (FAIR) Act, known in an earlier version as the Freedom to Farm Act, was signed into law by President Clinton, significantly changing U.S. agricultural policy. Because of the watershed changes, the FAIR Act required 14 months of debate to complete the legislative process. The most important features of the FAIR Act are the cut in government spending on farm price-income support, and the provision to allow farmers virtual production flexibility without limiting the type of crops that can be planted or requiring that acreage be idled. The act represents the first major reform of U.S. agricultural policy since Franklin Roosevelt’s New Deal, and is also the most controversial piece of agriculture legislation since the Great Depression. Experts disagree concerning what impact the FAIR Act will have on U.S. agriculture.

As Schwert argued in 1981, stock market data measure the impact of regulatory changes better than other data because asset prices determined in the stock market incorporate all relevant information as soon as it is available. Hence, a growing literature uses stock market data to explore the consequences of regulatory changes. However, there are only a few studies that have used stock market data, or event-study methodology, to empirically examine agriculture-related issues. These include three works by Johnson, Mittelhammer, and Blayney (1991, 1992, and 1994) for the tobacco industry, for the pesticide industry, and for the meat packing industry, respectively.

The main objective of this study is to examine how the passage of the FAIR Act affects the major agricultural input suppliers, using a variant of event-study methodology and stock market data. In examining the influence of agricultural policy on American agriculture production, it is difficult to use stock market data directly, because the agricultural production firms that are publicly traded are normally a special type, with none of them producing major crops such as wheat or cotton. However, farm laws usually focus on commodity programs, and thus major crops may play a particularly important role in these event studies. Conversely, this study represents well the agricultural chemical and farm machinery industries, because most of the larger firms comprising these input suppliers for agriculture are publicly traded. On the other hand, their market response to an agricultural policy also may be a good proxy for policy impact on agricultural production since their financial prospects are positively correlated with demand from agricultural production.

Additionally, our flexible event-study model will help test the timing of market response to the legislation. Johnson, Mittelhammer, and Blayney (1991) stated that the introduction of legislation may provide more information to the market than the passage of legislation, which is particularly true for agricultural policy. Their argument was based on the precedent (at the time of their study) that no U.S. President had vetoed any major agricultural policy legislation in the last decade. In a study on stock market response to 1986 tax law, Jang argued that the investors’ revision of stock valuation may have started around the date of the Senate Finance Committee proposal, and was virtually completed at the time of conference approval. Because the 1995 farm bill (an earlier version of the 1996 FAIR Act) was vetoed, we are afforded an opportunity to test the possible change in timing of significant market response.

The remainder of this article is structured as follows. First, we provide a review of the main features of the FAIR Act and its likely impact on the two major agricultural input industries, followed by a discussion of event-study methodology as it relates to regulations. Next, we describe the data, the empirical model, and the tests of the hypotheses. The empirical results are then reported and conclusions are presented.

Agricultural Policy Reform and the FAIR Act <top>

U.S. farm programs originated from President Franklin D. Roosevelt’s New Deal, when protective federal intervention was considered necessary to stabilize farm income. The Agricultural Adjustment Act of 1938 established, for the first time, the basic price-support and production control system for nonperishable agricultural commodities. This law remained in existence for more than 50 years. The Agricultural Act of 1949, the last permanent agricultural legislation, gave the Secretary of Agriculture more flexibility in setting price-support levels. By the mid-1980s, farm program policies faced increasing criticism, largely due to the huge and rapidly growing implementation cost. For example, costs under the 1981 farm bill were projected to be about $11 billion over its four-year duration, but a devastating farm depression forced the government to pay $54.7 billion to buy huge amounts of surplus commodities and to support farmers’ incomes with artificially high prices.

In 1985, Congress began its first attempt at substantial reform. The 1985 farm bill—the Food Security Act of 1985 (PL 99-198)—lowered artificially high prices but retained a lifeline to farmers through massive income-support payments. Although the cost ($88.6 billion over five years) was very high, the bill did more to nurse U.S. agriculture back to health than any collection of government programs since the Great Depression. The Food,

Agriculture, Conservation, and Trade Act (FACTA) of 1990 continued the reform. FACTA froze price and income-support rates at existing levels, though it did not reduce subsidy levels.

However, the above changes in farm policy generally occurred in relatively small increments. In contrast, the new farm bill, which was signed into law as the 1996 FAIR Act, represents a major change in U.S. farm policy (Knutson, Smith, Outlaw, and Woods). The major reforms in farm policy embodied in the FAIR Act can be categorized into the areas of (a) target price, (b) production control, (c) nonrecourse loans, and (d) the acreage reduction requirement.

The FAIR Act eliminates the target price, or deficiency payments provision. The replacement is a set of seven-year decoupled "transition" payments, which guarantee a 7% annual fixed but declining payment over the period 1996­2002 and total $35.7 billion. The term "decoupled" means that the size of the payment does not depend on the amount of crop produced or the level of the market price. Farmers are given greater flexibility to make planting decisions with the elimination of annual acreage-idling programs and with the freedom to plant any crop in contract acres, excluding fruits and vegetables. Meanwhile, farmers can receive transition payments. Previously, the government required farmers, including corn and soybean growers, to plant the same crops year after year in order to continue receiving subsidies. The basic nonrecourse commodity loans are retained, although in a modified form that combines the language of the nonrecourse loan with that of the marketing loan. Under the relevant provisions, farmers may receive a loan from the government at a designated rate per unit (loan rate) by pledging and storing a quantity of a commodity as collateral.

Obviously, the 1996 FAIR Act will force U.S. agricultural production to become more fully driven by market forces. The major advantage of the market-oriented farm policy is that agricultural prices and resource allocation are far less distorted. Nevertheless, there is controversy over the economic impact of the FAIR Act. Some people are concerned about the greater price risk that farmers will face. One serious widespread concern is that producers will bear greater income risk because the effect of price fluctuation on farmers’ income will not be completely offset by the fixed and declining government subsidies under the new law (Young and Shields; Knutson et al.). Moderate size farms, which already have problems competing, will face an even greater challenge (Knutson et al.).

A positive economic effect of the FAIR Act on U.S. agricultural production may be expected for several reasons. First, while the reduction of farm supports primarily hurts the interests of large agricultural producers, these producers may have a more stable financial situation for surviving the shock. The historic justification of farm programs is based on income equity, but that argument is no longer valid because the per capita income for commercial farmers with gross sales over $100,000 is not lower than the per capita income of nonfarmers (Flinchbaugh and Edelman).

Second, farmland values and U.S. exports are also likely to be higher in the long term (Young and Westcott). Trade liberalization and a more open world market supported by recent NAFTA and Uruguay Round Agreement of GATT bring potential for at least a partial substitute of foreign demand for government payments in the farm income stream (Flinchbaugh and Edelman). High price supports and mandatory production controls do not fit with a policy of expanding world trade. And finally, U.S. agricultural production is competitive, as evidenced by an increase over time of crop yields and a decrease in the cost of production—both measures of productivity.

Hence, the FAIR Act may expand U.S. agricultural production or induce farmers to plant more profitable high value-added products. In the context of our focus, both cases probably will increase demand for agricultural inputs and improve the financial prospect of agricultural input industries. For purposes of this study, the agricultural chemical industry is characterized as a variable input supplier, while the farm machinery industry is viewed as a quasi-fixed input supplier. Thus, the impact of FAIR may be different on both industries; i.e., the agricultural chemical industry may be affected more by the act than the farm machinery industry. Also, the FAIR Act does not repeal the Agricultural Act of 1949, and the duration of FAIR is only seven years. The implementation of FAIR still depends on unpredictable future events. Hence, the short-term effect may be more significant than the long-term effect.

In the following sections, the impact of the FAIR Act on agricultural input suppliers will be examined using information from the stock market.

Regulatory Event-Study Methodology <top>

Standard event-study methodology is based on a semi-strong form of financial market efficiency. The finance literature supports a semi-strong form of efficiency in the capital market, particularly in the stock market. Semi-strong efficiency requires stock prices to reflect all publicly available information.

Three commonly used methods in the literature are the mean-adjusted model, the market-adjusted model, and the market risk-adjusted model or market model. Brown and Warner (1980, 1985) provide valuable insights on the selection of these three methods. According to their simulation results, at least one of these methods, usually the market risk-adjusted model, performed as well as other more complicated approaches under most circumstances.

However, event studies on regulation or government policy topics require special treatment. Normally, regulatory event studies are involved with event date clustering; that is, the regulation is in effect for all relevant firms on the same day. If the focus is on a particular industry, the study also must address industry clustering, which is closely related to size effects.

Event date clustering can lower the number of securities whose event periods are independent. Hence, such clustering increases the variance of the performance measure (abnormal returns), and lowers the power of the test to detect abnormal performance (Brown and Warner 1980). Binder (1985), among others, suggests using a multivariate regression model (MVRM) [or joint generalized least squares (JGLS)] to correct the cross-sectional variance dependence. In a regulatory event study, where all events are aligned in calendar time, the explanatory variables are identical for each sample firm, and MVRM parameter estimates are identical to those obtained using ordinary least squares (OLS). The advantages of MVRM, as noted by Binder, include explicit incorporation of contemporaneous dependence of the disturbances and heteroskedasticity across the equations, and flexibility in testing joint hypotheses.

In the case of cross-sectional dependence, if it is not too severe, the standard OLS estimators actually may have smaller sampling variances (Zellner). Many previous event studies simply used OLS. Malatesta, following Brown and Warner’s simulation strategy of random selection of sample firms, demonstrated that the MVRM abnormal performance estimator is equivalent to the OLS estimator. More recently, Ingram and Ingram emphasized that the problem of cross-sectional dependence is exacerbated if all of the firms in the sample are in the same industry. Their Monte Carlo simulation results convincingly show that MVRM or JGLS is superior over OLS in the case of regulatory event studies characterized by both event clustering and industry clustering.

The MVRM procedure does have some limitations, however (Binder 1985). The distributions of test statistics available (LM, LR, W, and F) in MVRM are known to be only asymptotically efficient. In a small sample, the bias seems to increase with the number of equations. The F- and Wald statistics approximate the empirical distribution well when there are fewer than 15 equations, but not when there are more than 15.

The size of firms influences regulatory event studies, particularly when the studies are involved with industry clustering. The size effect can adversely influence an event study in two ways. First, if all firms within an industry are size clustered, maintaining a very similar size, the use of a market index to estimate abnormal returns may cause biases. The use of a market index in standard event-study methodology is justified when the abnormal returns of component securities are subject to normal distribution with a mean of zero, when no new marketwide information arrives. The existence of a size effect on securities returns can make the selected securities in the study always have abnormal returns, as measured by market portfolio performance.

Dimson and Marsh argue that performance measures in event studies are seriously biased when: (a) the measurement interval is long, (b) event securities differ systematically in size or weight from the index constituents, and (c) the size effect is large and/or volatile. The biases are likely to be greater when CAPM-type methodologies are used.

Kothari and Wasley assessed the impact of firm size clustering on test statistics based on market-adjusted and market model abnormal returns. Their simulation results suggest that when event dates are clustered in calendar time and events affect only large or small firms, conventional t-tests using market-adjusted or market model measures of abnormal performance are specified incorrectly.

James provides a discussion of the second way that failure to address the size effect can misrepresent the results of an event study. Studies that do not explicitly control for the size effect may fail to capture the different response of firms of distinct size to regulatory change. He stresses that if empirical studies examining the impact of regulatory change on firms at the industry level are done correctly, they may show important differences among classes of firms within the industry.

Based on the above considerations, we argue that regulatory event studies should allow for the problems associated with event date clustering, industry clustering, and size effect. As demonstrated in the following section, the combination of portfolio construction and the MVRM model serves to address these concerns.

Data and Empirical Methodology Data and Events <top>

The data for this analysis consist of daily stock returns for two specific input supplier industries listed on the New York Stock Exchange (NYSE), the American Stock Exchange (ASE), or the over-the-counter (OTC) market between December 12, 1994 (when consideration of the farm bill was first announced) and April 4, 1996 (when the FAIR Act was signed into law). The two industries are the agricultural chemical industry [Standard Industrial Classification (SIC) code 287] and the farm machinery industry (SIC code 3532). This actually represents the largest sample to serve our purposes, because only specific input suppliers may be affected by agricultural policy changes to a similar degree.

The firms in the final sample were traded during the entire study period. Daily return data were collected from Datastream databank. None of these firms had merger or failure activity during the estimation or event period used in the study. (A listing of these firms is provided in the Appendix.) The portfolios were constructed by grouping the firms according to their asset size, using a cutoff point of $100 million. As a check, the groups were categorized by their market value of equity. The new criterion did not change the groupings.¹

^{1Two reviewers suggested moving Allied Products Corp. into the small machinery firms portfolio. We found that the reported empirical results were qualitatively unchanged.}

Finding an appropriate event period was very important for investigating the market reactions to the new farm bill. The FAIR Act of 1996 experienced the longest farm bill debate in U.S. history and was frequently revised throughout the process. Consequently, as investors continually reassessed the probability of a farm bill proposal becoming law, they revised security values accordingly. Hence, the event period of the new farm bill should be sufficiently long to capture the full effect of the change, but short enough to reduce noise. Careful examination of the FAIR Act’s legislative process helped determine the event period.

The FAIR Act had two rounds of enactment. Early in February 1995, a congressional committee began conducting hearings on the 1995 farm bill, an earlier version of the FAIR Act that was scheduled to be completed during 1995. However, in December 1995, President Clinton vetoed the budget reconciliation package, citing the 1995 farm bill as a veto message. Congress went through another round of enacting the farm bill, i.e., the 1996 farm bill (the FAIR Act). That bill was signed by President Clinton into law on April 4, 1996. Actually, these two farm bills were quite similar, with only a few changes represented by the final 1996 version.

A careful search of the Wall Street Journal, the Wall Street Journal Index, and the Congressional Quarterly Almanac provided important event dates on which new information concerning the regulation (the FAIR Act) became publicly available. The empirical results reported here are based on the use of announcement days to determine event dates. We also allowed one more day for the market to disseminate information, and defined each event day as two-day windows. The results were the same qualitatively, and are not reported here.

Table 1 describes the major events surrounding the passage of the FAIR Act. Care was taken to make certain the events chosen for the test were unanticipated by the markets. The Wall Street Journal Index and database search results were used to ensure that no other announcements which directly affected the agricultural sector or these input suppliers occurred on the same event dates.

Methodology <top>

The impact of regulatory reform on major agricultural input suppliers was estimated with a multivariate regression model (MVRM) similar to the methodology used in several earlier studies (Schipper and Thompson; Binder 1988; Allen and Wilhelm; Cornett and Tehranian; Clark and Perfect; Liang, Mohanty, and Song; Wagster). In fact, MVRM has been largely accepted in event studies on regulations and policy (Binder 1988; Liang, Mohanty, and Song; Wagster). The impact of the FAIR Act was estimated with an MVRM following Binder (1985). Each firm was categorized by size and by industry into one of four portfolios: (a) large agricultural chemical firms, (b) small agricultural chemical firms, (c) large farm machinery firms, and (d) small farm machinery firms.  Firms reporting assets valued above $100 million (at the beginning of the study period) were considered large firms.

Using the MVRM methodology, the abnormal returns (γ_it) were parameterized in the individual return equations:

(1)

where R_it is the daily rate of return (defined as the ratio of price change between day t and t - 1, plus dividend on day t, over price in day t - 1) on the ith size/industry portfolio (where i = 1 for large chemical, 2 for small chemical, 3 for large machinery, and 4 for small machinery) on day t (where t = 344 daily observations from December 12, 1994 to April 4, 1996); R_m,t is the daily rate of return on the S&P 500 Index on day t; α_i is an intercept coefficient for the ith portfolio; β_1i, β_2i, and β_3i are lag, contemporaneous, and lead coefficients, respectively, of R_m,t for the ith portfolio; γ_ia is the effect of the ath regulatory announcement on the ith portfolio; A is the total number of regulatory announcements; D_a is a dummy variable for each regulatory announcement (equal to one on the day of the ath announcement and zero otherwise); and u_it is an error term for the ith portfolio on day t. The model assumed that its parameters were stationary over the analysis period and that the residuals were independent and identically distributed within each equation.

The standard hypotheses about average or cumulative average abnormal returns in typical event-study methodology, as well as more general hypotheses, can be tested within the above MVRM framework. Specifically, we formulated and tested the wealth effect of the FAIR Act for each portfolio on each event day, the overall wealth effect of the FAIR Act for all portfolios across all event days, the wealth effects of the FAIR Act on size/industry portfolios, and the wealth effects of the FAIR Act on the agricultural chemical industry and the farm machinery industry.

Empirical Results <top>

In this section, the results of the hypothesis tests are reported. The MVRM regression was run and then four hypotheses were tested. The test results show no heteroskedasticity or autocorrelation problems for any OLS regression in the system of regressions. The goodness-of-fit coefficient (R²) is 0.35 for the whole system of four regressions. Table 2 reports the regression results for the seven large agricultural chemical firms, eight small agricultural chemical firms, six large farm machinery firms, and four small farm machinery firms comprising our sample. The market beta coefficients (β_2i) are all positive and less than one, and all are significant at the 5% level with the exception of one market beta coefficient in the small machinery firm portfolio. The results were also robust for alternative starting points of the study period.²

²One reviewer suggested using more observations belonging to the period prior to FAIR-related news. Thus, we tested three alternative starting points— 06/01/94, 06/27/94, and 08/22/94—corresponding to 138, 120, and 80 additional observations, respectively, before the original starting point (12/12/94). We ran the MVRM regression on the three extended study periods, i.e., 06/01/94 to 04/04/96 (omitting two sample firms without complete data), 06/27/94 to 04/04/96 (omitting one sample firm without complete data), and 08/22/94 to 04/04/96. The results remained qualitatively the same.

It is also worth noting the gains in estimation efficiency provided by the MVRM methodology used in this study (except for the flexibility in hypothesis testing). The Breusch-Pagan LM test statistics are 13.31 with six degrees of freedom (Breusch and Pagan). The relevant critical value of χ² is 12.59 at the 5% significance level. Thus, diagonality of the covariance matrix can be rejected at 5%, and the existence of error contemp-oraneous correlation guarantees MVRM estimation to be at least as efficient as the OLS estimation. This is consistent with the findings reported by Ingram and Ingram.

Wealth Effect of FAIR on Each Event Day <top>

The first hypothesis tested was H_o: γ_ia = 0, where i = 1, 2, 3, 4, and a = 1, 2, ..., 13 (i.e., the abnormal return for a specific portfolio equals zero on a specific event day). The purpose of each test was to determine the wealth effects on each portfolio resulting from each regulatory announcement. The average abnormal returns (γ_ia = 0) and the t-statistics for each of the 13 events across the size/industry portfolios are presented in Table 2.

As shown in Table 2, no clear pattern of abnormal returns can be determined. Small farm machinery firms received a 5.6 percentage point abnormal return when Rep. Lugar proposed a plan to cut farm spending by $15 billion over five years (event 2). Small agricultural chemical and small farm machinery firms both received abnormal returns when the administration proposed cutting farm programs by $5 billion over seven years (event 6). This was one day after President Clinton vetoed the proposed 1995 farm bill. Stocks of the large agricultural chemical and large farm machinery firms were not significantly affected until after the House and Senate conference approved the compromised 1996 farm bill.

Stocks of the large agricultural chemical firm portfolio had significant abnormal returns on the day of conference approval (event 10). This is consistent with the stock market response from legislative events suggested by Johnson, Mittelhammer, and Blayney (1991), and by Jang. However, stocks of three portfolios (large agricultural chemical firms, and both large and small farm machinery firms) reacted significantly to the last two events in the legislative process—the passage of the FAIR Act by the House (event 12), and the signing of the FAIR Act into law by the President (event 13). This pattern of additional stock adjustments after the day of conference approval was different than the pattern suggested by Johnson, Mittelhammer, and Blayney (1991), and by Jang. The new stock-adjustment pattern reported here may have been caused by the President’s previous veto.

Surprisingly, large and small farm machinery firms both experienced negative abnormal returns after the House passed the 1996 farm bill (event 12) and the President signed the 1996 farm bill into law (event 13), respectively. It is possible the market determined that the potentially increased risk associated with the FAIR Act might lead to increased business risk, thus decreasing the demand for farm machinery. However, it is not clear from looking at the individual event dates what overall impact the FAIR Act had on the agricultural chemical and farm machinery firms.

Overall Wealth Effect of FAIR Across All Event Days <top>

The second hypothesis tested was H_o: γ_ia = 0 " i,a (i.e., the abnormal returns for each portfolio are jointly equal to zero). The test for this hypothesis was conducted by examining all announcement days to determine whether there were jointly nonzero abnormal returns among all event dates for every portfolio. This was done using a Wald test, producing a Wald statistic of 73.01. Compared to the critical value, χ² (52, 0.05) = 70.92, the test statistics suggest that the null hypothesis should be barely rejected at the 5% significance level. This test provides evidence that there were significant abnormal returns existing among the 52 total announcements.

In the following section, we examine the impact of the 13 separate announcements comprising our event dates and the effect of these announcements on the large and small firms within the two industries.

Wealth Effects of FAIR on Size/Industry Portfolios <top>

The third hypothesis tested was H_o: where i = 1, 2, 3, 4 (i.e., the abnormal returns for each of the size/industry portfolios are equal to zero over all announcements). This hypothesis was tested using four Wald tests (one for each portfolio). Table 3 reports the results of these tests. No significant cumulative abnormal returns were found for either large or small farm machinery firms. This implies that the wealth of large and small farm machinery companies did not change due to the passage of the FAIR Act.

We did find, however, that large agricultural chemical firms received significant positive cumulative-abnormal returns (Table 3). Positive cumulative-abnormal returns also appeared in the case of the small agricultural chemical firms, but were insignificant. This finding reveals that large agricultural chemical firms may benefit more from the FAIR Act than smaller agricultural chemical companies. A plausible explanation is that the production of some crops may actually decrease, while the overall production in agriculture may increase. Local or smaller agricultural chemical firms could be hurt because of shifts in production. Larger agricultural chemical companies with greater diversification and larger geographical areas might benefit from the increase or shift in production, net of some possible loss also caused by production shifts.

Wealth Effects of FAIR on Each Agricultural Input Industry <top>

The final hypotheses tested were H_o:

(i.e., the abnormal returns for each portfolio in the agricultural chemical industry are jointly equal to zero over all announcements); and H_o:(i.e., the abnormal returns for each portfolio in the farm machinery industry are jointly equal to zero over all announcements). Two Wald tests were performed (one for each industry). The tests provided valuable information about the cumulative wealth effects or overall economic impact of the FAIR Act on each agricultural input industry.

The alternative hypothesis here is that the two industries are expected to have positive abnormal returns, and thus the one-tailed test is appropriate. The Wald test statistic is 2.08 for the farm machinery industry. The null hypothesis could not be rejected at either the 5% or 10% significance levels. This is not surprising given the results found for the third hypothesis (Table 3). The Wald test statistic is 5.65 for the agricultural chemical industry; thus, the null hypothesis can be rejected at the 10% significance level and can almost be rejected at the 5% level (p-value = 0.059).

We therefore conclude that the passage of the FAIR Act has had an economic impact on the agricultural chemical industry, but not on the farm machinery industry. A plausible explanation for this finding is that the market expects an increase of agricultural production or a shift to more profitable crops. This would immediately increase the demand for variable inputs, such as fertilizer and other products, produced by agricultural chemical firms, but not for quasi-fixed inputs such as farm machinery.

Conclusions <top>

In this study we investigated the effect of the passage of the 1996 FAIR Act on returns to shareholders of two major agricultural input industries. The FAIR Act contains provisions to make American agricultural production more market oriented. The two most important provisions are (a) the gradual reduction of federal income support to agricultural production over the next seven years, and (b) the greatly enhanced production flexibility. The FAIR Act is intended to lower the burden on taxpayers who subsidize agriculture while maintaining a smooth transition for the agricultural sector.

To isolate the effect of the FAIR Act on large firms and small firms, the sample was divided into four subgroups: large agricultural chemical firms, small agricultural chemical firms, large farm machinery firms, and small farm machinery firms. The FAIR Act not only may affect the firms in different industries, but also firms of different sizes in different and predictable ways. This is due to a difference in competitive advantages between large and small firms.

The results suggest shareholders of large agricultural chemical firms have benefitted from the enactment of the FAIR Act. It is possible that the market expects an increase in agricultural production or a shift to more profitable crops, both of which tend to increase demand for more specific inputs. The market expects variable input suppliers (agricultural chemical firms) but not the quasi-fixed input suppliers (farm machinery firms) to benefit significantly from the enactment of the FAIR Act. Furthermore, the market expects that the large firms in the agricultural chemical industry will benefit, while the smaller agricultural chemical firms will not.

References <top>

Allen, P.R., and W.J. Wilhelm. "The Impact of the 1980 Depository Institution Deregulation and Monetary Control Act on Market Value and Risk: Evidence from the Capital Market." J. Money, Credit, and Banking 20(1988): 364­80.

Binder, J.J. "On the Use of the Multi-variate Regression Model in Event Studies." J. Accounting Res. 23(1985): 370­83.

———. "The Sherman Antitrust Act and the Railroad Cartels." J. Law and Econ. 31(1988):443­68.

Breusch, T.S., and A.R. Pagan. "The Lagrange Multiplier Test and Its Applications to Model Specification in Econometrics." Rev. Econ. Stud. 47 (1980):239­54.

Brown, S.J., and J.B. Warner. "Measuring Security Price Performance." J. Fin. Econ. 8(1980):205­58.

———. "Using Daily Stock Returns: The Case of Event Studies." J. Fin. Econ. 14(1985):3­31.

Clark, J.A., and S.B. Perfect. "The Economic Effects of Client Losses on OTC Bank Derivative Dealers: Evidence from the Capital Market." J. Money, Credit, and Banking 28 (1996):527­45.

Cornett, M.M., and H. Tehranian. "An Examination of the Impact of the Garn-St. Germain Depository Institutions Act of 1982 on Commercial Banks and Savings and Loans." J. Finance 45(1990):95­111.

Dimson, E., and P. Marsh. "Event Study Methodologies and the Size Effect." J. Fin. Econ. 17(1986):113­42.

Flinchbaugh, B.L., and M.A. Edelman. "The Political Economy for the 1995 Farm Bill." In 1995 Farm Bill: Policy Options and Consequences, edited by R.D. Knutson. Texas Agricultural Extension Service, Texas A&M University System, 1995.

Ingram, M.A., and V.C. Ingram. "Consistent Estimation of Residual Variance in Regulatory Event Studies." J. Fin. Res. 16(1993):151­60.

James, C.M. "An Analysis of Intra-Industry Differences in the Effect of Regulation: The Case of Deposit Rate Ceilings." J. Monetary Econ. 12(1983): 417­32.

Jang, H.-Y. J. "The Market Reaction to the 1986 Tax Overhaul: A Study of the Capital Gains Tax Change." J. Bus. Fin. and Accounting 21(1994): 1179­93.

Johnson, M.S., R.C. Mittelhammer, and D.P. Blayney. "The Impact of Regulation on Shareholder Wealth in the Tobacco Industry: An Event-Study Approach." Agr. Fin. Rev. 51(1991):21­34.

———. "Pesticide Industry Regulation: Economic Consequences as Revealed Through Stock Price Behavior." Agribusiness 8(1992):347­64.

———. "Stock Price Reaction to Regulation in the Meat Packing Industry." J. Agr. Econ. Res. 45(1994):31­40.

Knutson, R.D., E.G. Smith, J.L. Outlaw, and W.F. Woods. "New Farm Bill: Watershed Change in Policy." White Paper, Agricultural and Food Policy Center, Texas A&M University, College Station, April 1996.

Kothari, S.P., and C.E. Wasley. "Measuring Security Price Performance in Size-Clustered Samples." Accounting Rev. 64(1989):228­49.

Liang, Y., S. Mohanty, and F. Song. "The Effect of the Federal Deposit Insurance Corporation Improvement Act of 1991 on Bank Stocks." J. Fin. Res. 21, no. 2 (1996):229­42.

Malatesta, P.H. "Measuring Abnormal Performance: The Event Parameter Approach Using Joint Generalized Least Squares." J. Finan. and Quant. Anal. 21(1986):27­38.

Schipper, K., and R. Thompson. "The Impact of Merger-Related Regulations on the Shareholders of Acquiring Firms." J. Accounting Res. 21(1983): 184­221.

Schwert, G.W. "Using Financial Data to Measure Effects of Regulation." J. Law and Econ. 24(1981):121­58.

Wagster, J.D. "Impact of the 1988 Basle Accord on International Banks." J. Finance 51(1996):1321­46.

Young, C.E., and P.C. Westcott. "1996 Farm Act Impacts: An Early Assessment." In Agricultural Outlook. U.S. Department of Agriculture/ Economic Research Service, Washington, DC, August 1996.

Young, E., and D.A. Shields. "Overview: FAIR Act Frames Farm Policy for Seven Years." In Agricultural Outlook. U.S. Department of Agriculture/Economic Research Service, Washington, DC, April 1996.

Zellner, A. "An Efficient Method of Estimating Seemingly Unrelated Regressions and Tests for Aggregation Bias." J. Amer. Statis. Assoc. 57(1962): 500­09.

Appendix <top>

Table A1 provides a listing of the seven large agricultural chemical firms, eight small agricultural chemical firms, six large farm machinery firms, and four small farm machinery firms comprising our study sample.