Transaction costs and the collection of information: Presale measurement on private timber sales

Journal of Law, Economics, and Organization

Volumn 16, Issue 1, 2000, Pages 166-188

  Leffler, Keith B ^a   Rucker, Randal R ^b   Munn, Ian A ^c  

^a UNIVERSITY OF WASHINGTON   (United States)

^b MONTANA STATE UNIVERSITY   (United States)

^c MISSISSIPPI STATE UNIVERSITY   (United States)

Author keywords

[No Author keywords available]

Indexed keywords

EID: 0034402694     PISSN: 87566222     EISSN: None     Source Type: Journal    
DOI: 10.1093/jleo/16.1.166     Document Type: Article

References (63)

1. Shelanski and Klein (1995) review the empirical transaction costs literature and provide an extensive list of citations.
2. Much of the research in the auction literature either ignores presale measurement costs or assumes such costs to be zero. This leads to a peculiar feature of many auction models - a competitive equilibrium that requires an infinite number of bidders.
3. The dataset used in the present study is not the same as that used in Leffler and Rucker (1991). Details on the collection of this new dataset are provided in Section 2 below and in Munn (1993).
4. See Thaler (1992) for a discussion of the history of the winner's curse in the context of auctions for oil drilling rights. French and McCormick (1984) present an intuitive discussion of Wilson's model.
5. As Matthews notes, similar results were part of the auction literature concerning offshore oil leases. That literature, however, was based on numerical simulations that did not provide any general comparative statics results. Further, the amount of information was not a choice variable.
6. To understand the intuition underlying this prediction in the context of our empirical analysis (timber sales), consider the following. Assume a tract whose expected value is $1000, and whose actual value may be either $900 or $1100 (that is, the probability density function of the tract's value is binomial). The potential gains from measurement arise from being able to identify (or increase the probability of identifying) the tract as either a high-value tract or a low-value tract. Contrast this tract with another tract whose expected value is $100,000 and whose actual value may be either $99,900 or $100,100. That is, the value density function of the second tract represents a variance preserving shift of the density function of the first tract. The incentives of risk-neutral buyers to measure, which arise from the possible differences in the tract value, are the same on the two tracts. Essentially, neither the marginal benefits nor the marginal costs of measuring have changed.
7. This framework is adapted from the graphical analysis first offered by Johnson (1979) and later extended by French and McCormick. See French and McCormick for an intuitive explanation for why the expected bid approaches the true value of the tract in this analysis as the number of bidders increases.
8. By ex ante uncertainty, we refer to the intrinsic uncertainty concerning the value of a good when no information on its characteristics is available. In our application to timber sales, the ex ante uncertainty would be the level of uncertainty prior to a formal timber cruise by either the seller (or a timber consultant acting as an agent for the seller) or prospective buyers. This construct corresponds to the initial level of uncertainty in theoretical models of behavior under uncertainty.
9. Note that this aspect of our graphical framework differs from French and McCormick, who held constant the level of individual measurement.
10. Intuitively, this shift occurs because, for a given number of bidders, an increase in the level of uncertainty leads to an increase in the difference between the highest and second-highest information samples (and all other ordered information samples). It is these differences that determine the level of the bid and the size of ΔX for any given number of bidders in this auction model.
11. 0*, which considered alone increases the prize to the winning bidder.
12. Private timber sale contracts are usually 24 months or less in duration. By defining our "period" as the typical contract length, we abstract from the temporal aspects of the cutting decision.
13. These methods are described in, for example, the discussion of cruising in Wenger (1984).
14. Alternatively, the volume of timber on a tract may be measured using a prism cruise, which typically is quicker than a fixed-radius plot cruise, but subject to greater variance. See, for example, Wenger (1984) for further details on the mechanics of this method. In Leffler, Rucker, and Munn (1996), we provide an analysis of the determinants of the choice of cruise method.
15. This variation in value is typified by the following examples of recent Appalachian hardwood market prices: The average price of first- and second-grade red oak is $1150/mbf as compared with $605/mbf for #2 common; first- and second-grade hard maple is $1645/mbf as compared with $605/mbf for #2 common. See Weekly Hardwood Review (April 24, 1998).
16. To see that an increase in volume increases variance, consider a tract with one tree whose perceived size and quality (according to the seller-provided information) are such that the expected value of the tract is $110, and that prospective buyers' perceptions are that the actual value of the tract is either $100 or $120, both with probability of .5. Now, suppose the number of trees on the tract doubles to two. Assuming that the expectations concerning the second tree are identical to the first (i.e., holding composition constant), the tract now has an expected value of $220 with potential realized values of $200 with probability .25, $220 with probability .50, and $240 with probability .25. The variance of the tract thus increases from 100 to 200 with the increase in volume. We discuss below the effects of an increase in value that can accompany an increase in volume or that can occur for other reasons.
17. This discussion implicitly defines measurement units in terms of the number of trees evaluated. In our discussion of the determinants of presale measurement in the previous section, however, we used time units to indicate the quantity of measurement effort. Using that convention, a denser timber tract increases the marginal benefit of measurement. Regardless of the semantic difference between time spent and dollars spent on measurement, the empirical implication (an increase in measurement) is unchanged. Further, such a shift in the marginal benefits of measurement increases the optimal "amount" of measurement, in terms of both the time spent measuring and the expenditures on measurement.
18. For a detailed discussion of the procedures used to collect these data, see Munn (1993). Because many timber sellers are small, nonindustrial, private landowners it was not cost effective to contact more than a small number of them.
19. Copies of the survey instrument are available on request from the authors. Note that each survey response provides us with information on either the buyer's cruising activities on individual tracts or the seller's activities, but not both.
20. The industry practice is to estimate amounts of different types of timber using different volume metrics (or log tallies). As examples, the volume of pine sawtimber is typically measured in thousand board feet (mbf) Scribner log tally, the volume of hardwood sawtimber is often measured in mbf Doyle, and the amount of pulpwood is often measured in cords. We obtain our estimate of the total volume of timber on each tract by using standard industry factors to convert all volumes to mbf Scribner log tally. See Timber Mart South, Avery and Burkhart (1983: 50-53), and Munn (1993: 18) for the conversion factors used and discussions of such conversions in the timber industry. Alternative empirical specifications (which require making fewer of the conversions described above), using the individual volumes of the separate species and commodity classes, yield results that are qualitatively the same as those shown in Table 3.
21. The industry practice is to estimate amounts of different types of timber using different volume metrics (or log tallies). As examples, the volume of pine sawtimber is typically measured in thousand board feet (mbf) Scribner log tally, the volume of hardwood sawtimber is often measured in mbf Doyle, and the amount of pulpwood is often measured in cords. We obtain our estimate of the total volume of timber on each tract by using standard industry factors to convert all volumes to mbf Scribner log tally. See Timber Mart South, Avery and Burkhart (1983: 50-53), and Munn (1993: 18) for the conversion factors used and discussions of such conversions in the timber industry. Alternative empirical specifications (which require making fewer of the conversions described above), using the individual volumes of the separate species and commodity classes, yield results that are qualitatively the same as those shown in Table 3.
22. As indicated above, we obtained information on the number of competing buyers for each sale in our sample from the buyers and seller/consultants who participated in our survey. This information is, of course, a part of the records kept by consultants. Moreover, buyers typically record the bids of competing bidders as they are opened at the auction. For the sales in our empirical analysis, we use this bid information to tally the number of bidders.
23. A change in harvesting costs causes a mean preserving shift in the distribution of value.
24. The principle underlying this measure of value comes from standard auction theory, which suggests that on average the winning bid approaches the true value of the tract as the number of bidders becomes large. Accordingly, we estimated a bid price equation that included the natural logarithm of the number of bidders as one of the explanatory variables. Other explanatory variables included the volumes of various species and classes of timber on the tract, the duration of the contract, access conditions to the tract, distance to mill, and dichotomous variables to distinguish salvage sales and sales with high-quality timber. We then estimated each tract's "true" value as the predicted value from this regression using the estimated coefficients, the tract's actual characteristics, and a "large" number of bidders (we obtained virtually identical results with both 10 and 20 bidders - the average number of bidders on the sales in our sample is 4.7 and the maximum is 11).
25. More specifically, the value of this variable for a given tract is the simple average of the prices of the furthest-out futures contract traded on three (randomly chosen) days from the middle of the quarter in which the tract was sold.
26. The standard deviation of FUTURES PRICE is about 5% of its mean. In contrast, the standard deviation of VALUE is about equal to the mean.
27. An additional comment concerns the industry practice of assessing cruising intensity on the basis of the percentage cruise. It might be argued that the quote in the text refers to predicted changes in this measure of cruise intensity rather than in the volume of timber cruised. Results similar to those reported in the text are obtained from a regression specification with the percentage cruise as the dependent variable - in particular, the estimated coefficients on the variables we include as proxies for a tract's value do not have a positive impact on the percentage cruise.
28. Because none of the 46 observations in this sample had zero seller measurement, we estimate this specification with OLS.
29. It might be expected that the level of seller measurement will be affected by the costs of cruising a tract as well as by the expected level of buyer measurement. When we included DENSITY as a proxy for the level of measurement cost, however, it was not significant.
30. We test this prediction, although our dataset includes only three per unit auction sales. Despite the small number of per unit sales, we found substantially lower cruising effort on per unit (as compared to lump sum) sales. The estimated coefficients on a dichotomous PER UNIT variable added to Regressions 1 or 2 in Table 3, for example, are significant at an α level of .10.
31. Leffler and Rucker analyze in greater detail these advantages, as well as the disadvantages of per unit sales.
32. Allen, Douglas W., and Dean Lueck. 1992. "Contract Choice in Modern Agriculture: Cropshare vs. Cash Rent," 25 Journal of Law and Economics 397-426.
33. _. 1995. "Risk, Uncertainty, and Contracts," Working Paper 173, Center for Studies in Law, Economics, and Public Policy, Yale Law School.
34. Avery, Thomas E., and Harold E. Burkhart. 1983. Forest Measurements, 3rd ed. New York: McGraw-Hill.
35. Barzel, Yoram. 1982. "Measurement Cost and the Organization of Markets," 25 Journal of Law and Economics 27-48.
36. Bulow, Jeremy, and Paul Klemperer. 1996. "Auctions Versus Negotiations," 86 American Economic Review 180-194.
37. Buyers of Forest Products in North Carolina. 1989. North Carolina Department of Natural and Community Development, Division of Forest Resources.
38. Cheung, Steven N. S. 1969. The Theory of Share Tenancy. Chicago: University of Chicago Press.
39. Crocker, K. J., and S. E. Masten. 1988. "Mitigating Contractual Hazards: Unilateral Options and Contract Length," 19 Rand Journal of Economics 327-343.
40. French, Kenneth R., and Robert E. McCormick. 1984. "Sealed Bids, Sunk Costs, and the Process of Competition," 57 Journal of Business 417-441.
41. Gaier, Eric M. 1995. "Endogenous Information Quality in Common Value Auctions," mimeo, Duke University.
42. Greene, William H. 1997. Econometric Analysis, 3rd ed. Upper Saddle River, NJ: Prentice Hall.
43. Hausch, Donald B., and Lode Li. 1993. "A Common Value Auction Model with Endogenous Entry and Information Acquisition," 3 Economic Theory 315-334.
44. Johnson, Ronald N. 1979. "Auction Markets, Bid Preparation Costs and Entrance Fees," 55 Land Economics 313-318.
45. Joskow, Paul L. 1987. "Contract Duration and Relationship-Specific Investments," 77 American Economic Review 168-185.
46. Kennedy, Peter. 1998. A Guide to Econometrics, 4th ed. Cambridge: MIT Press.
47. Kenney, Roy W., and Benjamin Klein. 1983. "The Economics of Block Booking," 26 Journal of Law and Economics 497-540.
48. Leffler, Keith B., and Randal R. Rucker. 1991. "Transaction Costs and the Efficient Organization of Production: A Study of Timber-Harvesting Contracts," 99 Journal of Political Economy 1060-1087.
49. _, and Ian A. Munn. 1996. "Transaction Costs and the Collection of Information: Presale Measurement on Private Timber Sales," unpublished manuscript, Montana State University.
50. Lyon, Thomas P., and Steven C. Hackett. 1993. "Bottlenecks and Governance Structure: Open Access and Long-Term Contracting in Natural Gas," 9 Journal of Law, Economics, & Organization 380-398.
51. Matthews, Steven. 1984. "Information Acquisition in Discriminatory Auctions," in M. Boyer and R. E. Kihlstrom, eds. Bayesian Models in Economic Theory. New York: Elsevier Science.
52. Milgrom, Paul. 1981. "Rational Expectations, Information Acquisition, and Competitive Bidding," 49 Econometrica 921-941.
53. _, and Robert J. Weber. 1982a. "The Theory of Auctions and Competitive Bidding," 50 Econometrica 1089-1122.
54. _. 1982b. "The Value of Information in a Sealed-Bid Auction," 10 Journal of Mathematical Economics 105-114.
55. Mulherin, J. Harold. 1986. "Complexity in Long-Term Contracts: An Analysis of Natural Gas Provisions," 2 Journal of Law, Economics, & Organization 105-118.
56. Munn, Ian A. 1993. "Forestry Consultants: Their Effect on Timber Sale Prices," Ph.D. dissertation, North Carolina State University.
57. Persico, Nicola. 1997. "Information Acquisition in Auctions," unpublished manuscript, UCLA.
58. Shelanski, Howard, A., and Peter G. Klein. 1995. "Empirical Research in Transaction Cost Economics: A Review and Assessment," 11 Journal of Law, Economics, & Organization 335-361.
59. Timber Mart-South. 1990-1993 (various issues). The Frank W. Norris Foundation, Warnell School of Forest Resources, University of Georgia.
60. Thaler, Richard H. 1992. The Winner's Curse: Paradoxes and Anomalies of Economic Life New York: Free Press.
61. Weekly Hardwood Review. 1997. Vol. 12, issue 25:9.
62. Wenger, Karl F., ed. 1984. Forestry Handbook, 2nd ed. New York: John Wiley & Sons.
63. Wilson, Robert. 1977. "A Bidding Model of Perfect Competition," 44 Review of Economic Studies 511-518.