estimated that, on the average, forestry assistance 
added $4,205 to the total sale price received by 
each landowner. 
Jackson discussed several implications of his 
study. First, nonassisted landowners tended to 
make high-grading selective cuts, leading to lower 
present values of future harvest yields. Based on the 
economic results, the private forestry assistance 
program could be expanded to provide positive 
economic returns. Also, using a price-prediction 
equation, Jackson found that economical sales 
could have volume as small as 35,000 board feet, 
translating into an area of 5 to 10 acres. He noted 
that small landowners might even be a logical group 
to receive private forestry assistance. 
Jackson (1985 unpubl.) recently concluded a 
study of the nontimber effects of assistance in Mon- 
tana. His preliminary results indicate no detectable 
difference in use of best management practices on 
forester-assisted and nonassisted ownerships, al- 
though the study did confirm the advantages of 
assistance in making timber sales and in encourag- 
ing good timber management practices. 
Straka and others (1986) recently performed an 
evaluation of the Mississippi Forestry Project. This 
project consisted of placing an extra State service 
forester in two areas in Mississippi in order to im- 
prove contacts with landowners and increase forest 
management practices. The new foresters contact- 
ed all owners in their area to promote increased 
management. Straka found that service-forester 
promotional activity and management assistance 
produced direct benefit-cost ratios of 20:1, 8:1, and 
3:1 and at real discount rates of 4, 7, and 10 percent, 
respectively. The program also generated large lo- 
cal economic impacts and increased tax returns. 
The study demonstrated that adding foresters and 
making intensive contacts with owners were prof- 
itable in the State. 
After reviewing the preceding studies and oth- 
ers on financial assistance, Royer (1985) described 
the preliminary results of a logit regression model he 
developed using data from an earlier survey of non- 
industrial private forest landowners in the South 
(Fecso and others 1982). He examined the refor- 
estation decision made by nonindustrial private 
owners as a function of tract ownership characteris- 
tics, personal characteristics, market variables, and 
public policy variables. Sixteen independent vari- 
ables were used in the analysis, divided into four 
groups. Owner variables included tract size, forest 
ownership as part of a farm, and the predominant 
local land use (urban, agricultural, mixed 
agriculture/forested, and forested). Personal char- 
acteristics included income, age, education, farm- 
ing as a primary occupation, absentee ownership, 
and plans to sell harvested land. Indices of 
stumpage prices for sawtimber and pulpwood, an 
index of reforestation costs, and advice by industry 
or consulting foresters constituted the market fac- 
tors analyzed. Financial and technical assistance 
(from FIP and private forestry assistance, respec- 
tively) were the relevant policy variables. 
A hierarchical statistical analysis of the data in- 
dicated that ownership variables alone would cor- 
rectly predict reforestation decisions only 17 per- 
cent of the time. Personal characteristics interacted 
with ownership variables, adding nothing to the 
model's explanation of reforestation probability. 
Economic (market) variables increased the model's 
probability of accurately predicting reforestation by 
13 percent. Public policy variables--the provision of 
FIP or public technical assistance--were most influ- 
ential, explaining 60 percent of these landowners' 
reforestation decisions. 
Royer then developed single-equation models 
that eliminated the effects of multicollinearity (inter- 
relatedness) among many of the independent vari- 
ables. This allowed interpretation of the effects of 
individual variables within each category. Partial 
derivatives and elasticities were calculated for each 
independent variable. Derivatives represented the 
"probability of reforestation corresponding to a one- 
unit increase in the independent variable evaluated 
at the means. The elasticity, which can be comput- 
ed only for continuous variables, reflects the per- 
cent change in the probability of reforestation corre- 
sponding to a percent change in the independent 
variable." 
Royer's results indicated that the asset posi- 
tions (income or forest ownership size) of landown- 
ers had a strong positive influence on the probabil- 
ity of reforestation. Pulpwood (but not sawtimber) 
prices had a positive but only modestly significant 
effect on reforestation decisions. Coefficients for 
technical assistance from both private and public 
foresters were positive and significant, as was the 
effect of public cost-sharing. Of the significant vari- 
ables, increases in reforestation probability, as indi- 
cated by the partial derivative, were greatest for the 
provision of public forestry assistance (about 66 
percent greater), followed by FIP expenditures (+50 
percent per dollar spent) and provision of private 
forestry assistance (+44 percent). Other statistically 
significant factors were much less influential: size 
(+0.04 percent per acre), income (+0.05 percent 
65 
