Historic, archived document

Do not assume content reflects current
scientific knowledge, policies, or practices.

——

pa

me : Ba iT OR

AAS
— Ss fi fist }
PO C4 NVC

‘ Flue-Cured Tobacco:

. Price Differences

Among lypes !1(a)-13

*
|
r
Marketing Research Report No. 9
}
i
>
United States Department of Agriculture
Bureau of Agricultural Economics
lf Washington 25, D. C.
>

| 1952

+
iy
i For sale by the Superintendent of Documents, U. S. Government Printing Office, Washington 25,D.C. - - Price 10 cents
e
I

bi

SOINONOODS TVHNLINDINDVY 4O NvVaNNA X-EleLy ‘OAN AYNLINOIYOV 4O LNAWLYvd3aa's'N

VWd WOd&d VLVd

pL GNV ‘EL ‘ZL ‘(Q)LL “(PJ LL S3dAL JO SVIYV
JLVWIXOYddV JLVDIGNI SINIT N3a¥OUPd GNV GITOS

NMOHS LON HOVI
SIYDV OOL NVHL SS371 HLIM SFILNNOD &S

1940 pud OOO'S | ie
666V71-000'8 Be
666'Z -O00'€
6667 -000| GF
666 -O0l

SddDV

(SISva LINM ALNNOD)

St76l
“"39VIUIV
ODDvdOl

qauno-3anis

ri

es

=

~

ilies

a.

— ont

—— —

FLUE-CURED TOBACCO: PRICE DIFFERENCES AMONG TYPES 11(a)-13 1/

By George Re Rockwell, Jr., Agricultural Economic Statistician

CONTENTS
Page
SUNN QacViMalerern sits eicveleieicie sic.s\sicie viele cioisieiele oe « slettelelsls cisis\s/ere\eis\se's\e\e ele iolejeje.s\e 6 0 s'e 1
Backes ouncumMace rd all: c.',/./s1.)svsloleie’s elsicieve oie/o e/cfaldielvicie eisis'ecosseces cine ees ceeeee 2
Purpose Of the. Study. coccccscccccccccccevcvcsccccccsccccvsccecccccccccee h
Season average prices SSCSSSCSHSHOHPSHSFTHSSSSSOHCSHSSSHRHFSSSSHSSDBSSSSHHOHTSESEHOOBLESE 5
Weekly average prices CHOCCHSEHSHCSOEHHOHOSCHOSEHOSCHOOHLOHSOCEHSCEHROCOHOE SHEL OELES 6

SUMMARY

Two aspects of flue-cured prices are given attention in this‘report. One
is the extent to which the average price of a group of United States grades
differs from one type of flue-cured tobacco to anothere These comparisons show
on the basis of a whole season the size and direction of price differences among
belts. As the price averages are for groups of grades identical in each belt,
it may be presumed that differences in price are not caused by differences in
quality. Such price comparisons for the 1945-9 seasons are shown in table 1,
page 6. The second aspect is the way in which the average price of a group of
U. Se grades of a single type varies during the course of a marketing season
and the factors that are apparently associated with this variation. This also
permits price comparisons among belts for various periods within a season. In
this connection, the influence on flue-cured prices of changing quantities of
certain qualities of leaf sold in the various flue-cured belts from week to week
in the course of the marketing season was investigated.

Flue-cured tobacco is grown in the coastal States from southern Georgia
and northern Florida as far north as central Virginia. Within this area are
five belts, each of which produces a type of flue-cured tobacco. Southern
Georgia and northern Florida produce type 1h; eastern South Carolina and south-
eastern North Carolina, type 13; eastern North Carolina, type 12; central North
Carolina, type 11(b); and the Piedmont section of North Carolina and Virginia,
type 1l(a).

The five types of flue-cured tobacco are graded at auction markets accord-
ing to Government quality standards. These standards classify tobacco of each
type into approximately 120 grade designations. It is, therefore, possible to
make a selection of grades for price comparisons among types.

Differences in price for the individual types of flue-cured tobacco are more —
accurately measured by comparing prices for an individual grade or a group of
identical grades of comparable quality than when over-all season averages for
individual types are compared with one another. When over-all season average

i/ The research on which this report is based was made under the authority of
the Agricultural Marketing Act of 196 (RMA, Title II).

-2-

prices are compared, differences among belts reflect differences in the average
quality of tobacco sold, as well as actual differences in price.

From 1945 to 1949, prices of selected grades2/ in type 11(b) or type 12
ranked highest most frequently, while those in type 1l(a) or type 13 most fre-
quently ranked lowest. This relationship of prices for the four types3/ of flue-
cured tobacco studied held for all quality groups for both the season as a whole
and for the weekly average prices during a season. :

Highest weekly prices occurred most frequently between the seventh and
fourteenth weeks during the marketing season for types 11(a)-13 and the lowest
either during the first 5 or the last 5 weeks of the season. Some have attributed
the low prices at the beginning or end of the marketing season to special con-
ditions during those parts of the season. It was thought that prices were low
at the beginning of the season possibly because purchasing companies had not yet
fully appraised the size and nature of the flue-cured crop and were low at the
end of the season because buyers made final adjustments in their purchases for
the season. An evaluation of the ranking of prices by types during the weeks
when all four types were marketed simultaneously showed conclusively that prices
during the beginning or the end of the marketing season were not lowest primarily
because of special conditions that affected those parts of the seasoned Other ‘
factors apparently were responsible for the price pattern that prevailed.

The proportion of total sales which selected grades accounted for in the
different belts varied considerably from one week to another. Apparently, how-
ever, differences in quantities of tobacco marketed in a single week did not
influence intertype price differences for that week for any of the groups of
selected grades.

Background Material2/

Flue-cured tobacco is grown in the coastal States from southern Georgia and
northern Florida as far north as central Virginia. This area is divided into
five belts, each of which produces a type of flue-cured tobacco. Type lh, Georgia ‘
flue-cured tobacco, is grown in southern Georgia and northern Florida; type 13,
South Carolina and Border North Carolina flue-cured, is produced in eastern South
Carolina and southeastern North Carolina; type 12, Eastern flue-cured, in eastern
North Carolina; type 11(b), Middle Belt flue-cured, in central North Carolina;
and type 1l(a), Old Belt flue-cured, in the Piedmont section of North Carolina
and Virginia.

1

2/ See notes on grade selections on page ll.

3/ Prices of types 11(a), 11(b), 12, and 13 tobacco were compared. Prices of
type 1h tobacco were omitted because it is marketed loose instead of tied in
hands. It is not practicable to separate the influence of the different form
in which type 1) tobacco is marketed from other factors that may operate to cause
it to sell consistently for less than the same grades of the other four flue-cured
typese

L/ See footnote 8/ on page 8.

5/ Background material prepared by Arthur G. Conover.

Flue-cured tobacco is the largest single class of tobacco grown in the

‘United States. The six classes of domestically produced tobacco--flue-cured,

fire-cured, air-cured, cigar filler, cigar binder, and cigar wrapper--include

a total of 2h types grown in the United States (25 if the Puerto Rican type is
included). Production of flue-cured tobacco in the years since World War II

has ranged between 55 and 63 percent of the tobacco produced in the United States.
During the five marketing seasons, 195 through 19,9, the total quantity of
flue=-cured tobacco auctioned each season averaged 1,200 million pounds, sub-
divided as follows: 257 million of type 11(a), 18 million of type 11(b),

396 million of type 12, 263 million of type 13, and 136 million of type 1h. This
huge volume of leaf went to auctions located in about 80 to 90 towns and cities
in the Atlantic seaboard States from Florida to Virginia. , The first flue-cured
marketings of a season start in the Georgia-Florida area in the last half of
July. By the latter half of September, the last group of markets in Virginia
and North Carolina begin sales. Flue-cured marketings are generally completed
by December. In some years, however, they continue into January, although
rarely into February.

In recent years, annual exports of flue-cured leaf have been substantial,
usually accounting for four-fifths of the total exports of unmanufactured tobacco
fran the United States. During the 195-9 marketing years, the average exports
of flue-cured tobacco totaled 46 million pounds (farm-sales weight) or 37 per-
cent of the average production for those years. From 195 to 1949, these exports
ranged between the low of 359 million pounds in 1947 and the high of 552 million
pounds in 1946. Approximately three-fourths of the flue-cured exports go to the
countries of western Europe. This includes the United Kingdom, which is the
leading importing country. The remaining one-fourth goes to numerous other
countries throughout the world. Grades of tobacco exported vary considerably by

country and, except for the United Kingdom and a few other countries, tend to

include grades of lower quality than those purchased for eventual consumption

in cigarettes in this country. Consequently, variations in export demand
probably affect somewhat price differences between grades and possibly between

types.

Flue-cured tobacco fram a current year's crop is not usable in manufacture
immediately following its purchase at auction. First, the tobacco is redried
and then prepared for storage. Storing the leaf for a considerable period is
an essential part of the processing operation before manufacture of the tobacco
product itself is begun. Aging of freshly cured leaf is necessary to remove
rawness or irritating properties and to develop aroma. The aging period varies
in length, but from 18 to 2h months are usually desirable. To insure con-

'Sistency of blend and uniformity of the final product, manufacturers must main-

tain large stocks of leaf tobacco of various types and grades from the crops of
several different seasons.

The size of carry-over stocks and the composition of their grades and types

_are among the factors that can influence the season average price received by

growers for the current crop of tobacco. In any single season, short-term

fluctuations in the volume of marketings would not be expected to affect market
price very much in view of the large inventories generally carried by manu-
facturers. During 1945-9, total stocks of flue-cured tobacco in the United
States on July 1 (just before the marketing season begins) ranged between 1.7 and
22 times the preceding year's domestic usings, or enough for a period of 20 to

oh

26 months. These total stocks were held in varying proportions by manufacturers,
dealers, or under Government loan through the growers! Flue-Cured Cooperative

pbapil eatin Corporation. Among the dealers were those who carried on af export
usiness.

Flue-cured tobacco is used primarily for the manufacture of cigarettes. It
has been estimated that in post-World War II years cigarettes have accounted for
around 95 percent of the total domestic use of flue-cured tobacco. The remainder
goes mostly into smoking mixtures and into plug chewing tobacco. Cigarettes
manufactured in this country characteristically include flue-cured, Burley,
Maryland, and imported tobacco, frequently referred to as "Turkish." Proportions
of each kind vary to some extent among manufacturers, but flue-cured tobacco
generally composes more of the average cigarette than the others. Leaf blends
of individual manufacturers are kept secret. However, it may be reasoned that
the substitutability of one kind for the other is kept within fairly narrow
limits, and that such classes of tobacco as fire-cured and dark air=-cured are
not at all substitutable for the kinds of cigarette tobacco mentioned above.
Among the flue=-cured types, it seems likely that closely similar grades may be
substitutable one for the other to a fairly wide extent. Blending leaf grades
and types to obtain a favorable combination of all quality elements in a uni-
formly finished product is a complex operation concerning which few, if any,
statistics are publicly available.

PURPOSE OF THE STUDY

Variations in market prices for the different flue-cured types of tobacco
have long been of interest to growers, warehousemen, and others connected with
the tobacco trade. Differences in prices among types of flue-cured tobacco have
been studied as part of an Agricultural Marketing Act of 196 (RMA, Title IT)
project in the Bureau of Agricultural Economics. If the reasons for these
differences in price could be explained, it might be possible for growers to
make adjustments in production or marketing practices that would equalize prices
among the flue-cured belts. If these changes brought about prices in all belts
equal to those in the belts presently commanding the highest prices, total
income from flue-cured tobacco would be increased. significantly.

This report is concerned with the influence of the quantity of leaf of
specified quality marketed upon differences in price among the flue-cured belts.
It answers the question as to whether prices of tobacco leaf of a certain
quality are depressed in one belt when marketings of that particular quality
of tobacco are large in relation to total marketings in that belt, and whether
prices tend to rise when marketings of that quality of tobacco are small in
relation to total marketings in the belt.

If it were found that the price of any quality of tobacco leaf is affected
by the proportion of total marketings which tobacco of this quality represents,
growers might be able to regulate the quality composition of their marketings
throughout the season so that the prices received during the season for this
quality of leaf would be more nearly stabilized. It would be expected that
this would yield growers a larger gross returne If this practice were followed
by all growers, it would be expected that differences in price still existing
between individual markets and between belts would be caused by other influences.

58

One phase of the study upon which this report is based was to measure
_differences in price more accurately by deriving prices for groups of identical
‘grades in each typee This involved ascertaining the extent and direction of
“differences in price among types 11(a)-13 flue-cured tobaceo for three quality
groups of leaf: better, medium, and poorer, for the seasons of 195 through
(1949. A second phase was to study the intertype price relations for individual
weeks within these seasons in order to establish more precisely the nature of
the differences. In this connection, the influence on price of proportions of
‘total marketings in the respective flue-cured belts accounted for by leaf of
4specified quality was investigated.

| Another phase of the study was to investigate whether differences in price
among types were related to the fact that marketing seasons in the different
belts were not exactly the same. Although considerable overlapping occurs when
‘auctions are going on simultaneously in all the belts for types 11(a)-13 flue-
4cured tobacco, type 13 markets open before the others and type 1l(a) markets
are the last to close.

| SEASON AVERAGE PRICESO/
A In making price comparisons, grades were assigned to three quality groups:
better, medium, and poorer. For the methods used in selecting grades, see notes
on page ll.

From 195 to 1949, type 11(b) season prices were highest for 3 of the
5 years within the better and medium grades, and type 12 prices were highest
for years within the poorer grades.

Lowest prices of the better group occurred in type 13 for 3 of the 5 years,
|lowest prices of the mediwn group were in type ll(a) for h years, and lowest
prices of the poorer group occurred in type 11l(a) for all 5 years.

4 In general, prices varied less from one type of flue-cured tobacco to
janother for better grades than for medium grades, and most for poorer grades.
This was true both in terms of cents per pound and in terms of percentages of
types 11(a)-13 average price for the respective quality groups. Compared with
types 11(a)-13 averages, the season average prices of better grades in any type
during these years differed from about 0.1 cent per pound to a maximum of about
3 cents, while the difference in season average prices of poorer grades amounted
to from 0.1 cent to about 3.5 cents per pound. AS a percentage of the types
11(a)-13 average prices for the respective quality groups, the largest differ-
ences in any type during these seasons amounted to h.5 percent in the better
grades and to almost 19 percent in the poorer grades.

Re These comparisons indicate that poorer grades are affected more than better
Brades by the factors that cause differences in price among typesof flue-cured
tobacco and that prices of type 11(a) and type 13 tobacco frequently are lower

|
]

6/ Season average price as used here is the simple average of season prices
of individual grades for each quality group in each type. Prices for individual
\grades are from the Production and Marketing Administration. They are based on
a sample of auction bids made throughout the season in representative markets.

= 6

than prices of type 11(b) and type 12 tobacco. As these averages are based on
whole seasons, prices for individual weeks within the seasons were examined in
order to determine how prices for a single type varied throughout the course of
a marketing season and how weekly prices of these four flue-cured types were
related.

Table 1.- Season average price per pound, flue-cured tobacco, types 11(a)-13
selected grades, and percentage comparisons, 195-9

: Average price per pound by type

Season

$ AS a percentage of average
of types 11(a)-131

me ee ee

:
Better grades

19L5 : 48.65 49.35 48.20 6.30 : 101.2 102.6 100.2 96.2
19h6 = 58.95 60.60 63.55 63.80: 95.5 98.2 103.0 103k
19h7 «57.90. 57.70. 57.05 57.10 s=200.8 100.h 99.3 99h
1948 =: 65625 66.10 65.75 65.25: 99.5 100.8 100.2 99.5
19h9 : 64.ehbO 64.95 6.05 63.20: 100.) 101.2 99.8 98.5
: Median grades
TONS 2273.89. hhc Hl) 78) c)l:c22 cameo. 2 99.54 <u. 101 See eutOgen
AG) = hell! 6.22) a; SO cin) 10h7.56) me Oreo 103.1 1OFS2 20137
I9NT 22 89° hy 89 367 2-2 522 “25 97.0 10.0 101.2 97.8
19185 3750.33 = Seely = 51522) 2051'.00: 2h 982 102.3 99.9 99.5
199) 4228.89 ~~ 52.33 9525001 116.78) :<45 9768 10h.7 10.0 93.6
s Poorer grades
1945 S36) 33015) pusteoie 13060) ; 90.1 96.9 107.9 105.1
196) 2218.02: (23657 23.96) 20.184 0968 109.6 111. 95 02
VON7 2293.02 -16507) 17695 2216575: see Gr.6 100.8 112.5 105.0
TOUS 218.934 2023s 27 B21 22h] ome 9 36 101.0 103.3 106.1
FOW9 es UNSB6 2 19677~ — 20609591768; Clee 108.0 114.2 96.6

1/season average price of each quality group of each type as a percentage of

types 11(a)-13 average of corresponding quality groupe

WEEKLY AVERAGE PRIcES?/

Weekly average prices of selected grades of flue-cured tobacco, types 11(a)-1
from 1945 to 1949 for each year except 1945 followed a general trend from low /
prices in the beginning of the marketing season, to high prices in the middle of |

/ Weekly average price is the simple average of weekly prices of individual
grades for each type in each quality group. For weeks in which prices were not ;
quoted for certain individual grades because they were not sold in sufficient
quantities, an average price for those grades was calculated and used in comput-
ing the average price of the quality group.

OO

=a

the season, to low prices toward the end of the season. Price ceilings were in
effect during the 195 season and prices did not follow the usual seasonal
pattern.

From 1946 through 1949, the high weekly price of each quality group occurred
between the seventh and fourteenth weeks of the marketing season for types 11(a)-13.
The low weekly price was reached for each quality group during the first 5 or the
last 5 weeks of the season, except for the medium group in 196.

During each year from 195 to 19h9, the highest weekly price in each quality
group occurred more often in type 12. For the better and medium groups, type 12
prices were highest in of the 5 years. In each of the 5 years, they were
highest for the poorer group.

Lowest weekly prices during these years occurred most frequently in type 11(a)
for the medium and poorer grades. They were lowest in 3 years for the medium
group and lowest in all 5 years for the poorer groupe The lowest weekly price
of the better group did not occur as consistently in any one type.

Price differences from the average of types 11(a)-13 for weeks during a
season when all of the four flue-cured types were marketed generally were larger
in each type for the poorer grades than for the better grades, both in absolute
terms and in relation to the average prices of types 11(a)-13. No consistency
was found for any one type in price differences from the averages for the various
weeks, either in absolute or relative terms. This indicates that the factors
that influence differences in price among flue-cured types affect the poorer
grades more than they affect the better grades.

Price Differences During Weeks When All Four Types Were Marketed.--A possible
explanation for the wide variations in weekly average prices is the time element.
For instance, prices of types 11(a) and 13 for all quality groups usually are
lower than prices of types 11(b) and 12, both for the season as a whole and for
weeks within a season. Lowest weekly prices usually are found at the beginning
or end of the season, at a time when only type 13 or type 11(a) markets are open,
or when the marketing of either types 11(b) and 12 are getting under way or are
being completed. The question arises as to whether type 13 low prices at the
beginning of the season result from incomplete knowledge on the part of purchas-
ing companies with respect to size and quality of the crop throughout the flue-
cured belt or whether they are due primarily to factors such as technical
differences between grades of type 13 tobacco and the same grades of other flue-
cured typese If incomplete knowledge of the size and quality of the whole flue-
cured crop is the important factor, buyers would be expected to bid conserva-
tively for the type 13 offerings during the first weeks of selling, until markets in
the type 12 district had opened and the size and nature of the majority of the
flue-cured crop had been more fully appraised.

With respect to the low prices of type 11(a) tobacco at the end of the season
when type 12 and 13 markets have closed and marketings in the type 11(b) district
are being completed, the question arises as to whether the low prices result from
final adjustments in purchases made by buyers for the season. In other words,
would the low prices at the end of the season be due to too heavy buying of other
types during the middle of the season with respect to requirements for leaf, and
therefore, less interest by purchasing companies in offerings at the end of the
season?

Ors

An examination of the prices paid for types 11(a) and 13 tobacco in relation
to prices paid for other flue-cured types during the middle of the marketing
season when all types sell at the same time should throw some light on these
questions.

From 1945 to 1949, all of these types were marketed simultaneously during

22 weeks, or fran 3 to 6 weeks in each season. If the low weekly price during
the season were due only to a seasonal factor, it would be expected that during
these weeks when all types were selling, type 11(b) or 12 prices would rank in
both first and second places and type 11(a) or 13 prices would rank in both
third and fourth places for a sixth of the time, or in between 3 and l of the

22 weeks. If the weekly prices ranked in this order for substantially more than
_k weeks, it would be assumed that some systematic factor operated to produce
differences in price between types 11(b) and 12 and types 11(a) and 13. In that
case, the low weekly prices of types 11(a) and 13 at the beginning and end of
the season would not be the lowest of the season primarily because of special
circumstances during those parts of the season.

: Ranking the weekly prices for the 22 weeks during the 5-year period when
all types were marketed simultaneously showed that the weekly price of either
type 11(b) or type 12 was in both first and second places and that the weekly
price of either type l1(a) or type 13 was in both third and fourth places for
13 weeks in the better group and for 15 weeks in the medium and poorer groupSe

This test provides strong supportS/ for the conclusion that low weekly
prices, which usually fall during the first or the last 5 weeks of marketing,
were not low only because of special circumstances at the beginning or end of
the marketing season for types ll(a)-13. It, therefore, suggests that other
factors operate to cause price differentials between types 1l(a) and 13 and
types 11(b) and 12.

Effects of Quantities of Selected Grades Marketed on Intertype Price
Differences.--A possible factor that might cause these intertype price differ-
ences for any quality group apparently was the proportion of total sales of a
particular type represented by sales of that quality group. For instance, if
sales of type ll(a) better grades were a relatively small part of the total
sales of type 1l(a) during a week compared to the proportion of total sales
of type 11(a) represented by better grades during the previous week, it would
be expected that the price of type 1l(a) better grades would be higher in the
second week, other things remaining the same. It would also be expected that
the price of better grades would vary from markets of one type to markets of
another type whenever the proportions of total sales in the respective types
accounted for by better grades differ widely.

This assumes that manufacturers are willing to pay higher prices for grades
of tobacco that happen to be scarce during a short period such as a week, so that

B/ If no systematic factors caused differences in the prices of these four
types, the chances of getting as great or more frequent ranking of prices of

type 11(b) or 12 in both first and second places and prices of type 1l(a) or 13
in both third and fourth places would be for the better group, about 1 in
100,000 repeated observations, and for the medium and poorer groups, about 1 in
10 million.

= 9a

the average quality of their purchases for the week would be approximately equal
to the quality of the standard flue-cured component used in the manufacture of
their product. If savings could be made in subsequent handling and sorting by

, acquiring the desired proportions of uniformly graded leaf, this would permit

payment of higher prices for grades that at any time were relatively scarce in
the markets of one of the belts.

Selected grades of each quality group accounted for widely different per-

centages of total sales in the respective types from one week to another and

from one type to another. For instance, the better group accounted for from
19 to 30 percent of the total sales of type 11(a) in the 1) weeks of the 197-9
marketing seasons in which all four types were sold. In the week ending
September 29, 199, better grades accounted for 20 percent of total type 11(a)
sales, 19 percent of total type 11(b) sales, 30 percent of total type 12 sales,

. and 25 percent of total type 13 sales.

It was thought, also, that buyers might regard the quality composition of
the marketings from week to week as an indication of the quality of the flue-
cured crop in each belt. So far as important buying companies revise their
estimates of the quality of the crop in any one of the flue-cured belts in the

light of changes in quality of its weekly marketings, prices in that belt would
_ tend to change inversely with changes in the quality of marketings. This

influence would probably be more important in the early part of the season.

In order to test these possibilities, correlations were run for each
quality group to test the relation between intertype relative prices and the
percentages of total sales the quality groups represented in their respective

types. Intertype price differences, represented by ratios of the price of one

type of the average price of types 11(a)-13, were used as the dependent variable.

For example, in attempting to explain the price differences between type 11(a)

and other flue-cured better grades, the dependent variable was the type 11(a)
better grades price divided by the types 11(a)-13 better grades average price,
and the four independent variables were the percentages of total sales of each

_ type accounted for by better grades.

The number of observations available was small. It was necessary to
estimate the weekly sales of selected grades producers. Recordsfor doing
this are available beginning with 19k7 only.2/ For 19k7 through 19h9, all

9/ There are no records of the number of pounds of individual grades of flue-

/ cured tobacco sold. It was, therefore, necessary to estimate the number of

pounds of selected grades for each quality group sold weekly in each of the
four types. These estimates were made from data reported in the Market News
Service sample.

Beginning with 1947, the number of pounds, as well as prices and number of

_lots, for each grade were reported in the sample. Estimates of the number of

pounds of selected grades sold weekly in all markets of each flue-cured type
_were made based on the number of pounds of selected grades in the sample during
the week. The percentage of total pounds in the sample that selected grades

constituted for each week was computed and this percentage was multiplied by

_ the total number of pounds of producers! sales in all markets of the type for

the week to obtain an estimate of the total number of pounds of selected grades

four types were sold simultaneously during 1) weeks. Correlations based on
these 1h observations failed to yield results that differed significantly from
zero. The coefficients of determination obtained ranged from .)\2 to .7he An
attempt was made to obtain significant results from the limited data available
by deriving the coefficient of multiple correlation from simple correlation
coefficients which were based partly on unmatched observations from the weeks
when fewer than four types were sold. This increased considerably the number
of observations on which some of the simple correlation coefficients were based,
but the results still did not differ ‘significantly from zero.

As a further check on the possible relation between weekly intertype price
differences and quantities of the various quality groups offered in the market,
the intertype variation in price was compared with the intertype variation in
supply. A coefficient of variation (F) of the intertype price ratios was com-
puted for each intertype comparison in price (type 1l(a) price/type 11(b) price,
etc.) for each quality group and compared with the corresponding supply coef-
ficient of variation. Two separate coefficients of variation in supply were
computed in each case. One was based on the ratio of percentages which the
quality group constitutes of total weekly marketings in each of the two types
concernede The other was based on the ratio of actual pounds of tobacco of the
quality group marketed in each of the two types concerned.

If a price coefficient of variation is absolutely small and if the corre-
sponding coefficient of variation in supply is large in relation to it, changes
in ratios of supply are not important causes of changes in price ratiose

Comparison of the price ratio coefficient with the supply ratio coefficient
shows, in the first case, whether intertype differences in price are closely
associated with changes in the percentages which are accounted for in total
sales of respective types by selected grades. It tends to show whether changes
in percentages which a quality group accounts for affect price, and might there-
fore, be an important cause of intertype differences in price. In the second
case, comparison of the price and supply coefficients of variation show whether
intertype differences in price are closely associated with changes in the actual
number of pounds of selected grades marketed in the types concerned.e Both com=
parisons tend to show whether the two types of flue-cured tobacco concerned are
competing products (that is, direct substitutes for each other) or whether they
are distinctly different products for which buyers will pay different prices.

If they are direct substitutes, their relative supply will affect their relative
prices very little. The converse, however, is not necessarily true in this
instance.

Price and supply coefficients of variation for types 11(a)-13 flue-cured
tobacco are shown in table 2. In all cases, price coefficients are absolutely
small and supply coefficients are large in relation to them. This indicates

sold in all markets of the type during the week. Weekly estimates of total sales
of selected grades were adjusted in line with estimated season total sales of
selected grades for each type. From these total figures, percentages which
selected grades of each quality group accounted for of total weekly sales for
each type were computed.

. Jae

that the proportion of total marketings in a type which a quality group consti-
_ tutes is not an important cause of intertype differences in price.
The intertype differences in price change most from week to week for poorer
grades and least for better grades, as indicated by the relative size of the eset
coefficients.

Table 2.- Coefficient of variation of price and supply
of flue-cured tobacco, by grade and type+

Coefficient of variation
Supply ratio of quality group to weekly
Type

: marketings based on-=
ratio : Percentages of Actual pounds

, respective types 8
Better grades

eo 002 08 88 e8
i
Q
i)

11(a)/11(b) : 0.8 14.6 2h.3
11(a)/12 H VA Ovel Ses
11(a) /13 : Led 38.7 106.2
11(b)/13 H 1.7 bloat 103.7
: Medium grades
¢ eee H 4.3 25.5 16.5
aie ine : 567 62.0 30.0
/13 3 5.8 64.5 128.0
tis) 22 ; hol 40.2 30.0
11(b)/13 : 6.0 2.8 115.2
12/13 ° he2 35.9 118. 0
2 Poorer grades
11(a)/11(b) : Jel MAC 30.0
11(a)/12 : 13.6 363 7.9
11(a)/13 : 11.0 31 190.0
11(b) /12 : 8.0 hh.O 40.3
11(b)/13 : 10.9 38.6 176.0
12/13 : 12.6 39.0 112.8

1/ Based on 1h weeks in the 19])7-h9 seasons when all of types 1l(a)-13 were

solde

| Notes on Grade Selection.-=-Better grades: Bll, BIF, B2L, B2F, B3L, HIF,

|) H2F, HOF, CLL, CoL, C3L, Cor, ClL, CULV, CLF, CSL, X1L, X1F, X2L, and X2F3
medium grades: BhFM, X3LV, X3FV, X3FM, "xhn, *xhLV, XLF, P3F, and PlL; and poorer
grades: BOR, BOGL, BOGF, B6GR, XSFM, PSL, PSF, P5G, NIL, N1R, N1G, N2L, NR,

and Ne.

- 12e=

Price was used to determine the appropriate quality group for a grade.
Grades were selected on the basis of the consistency of their prices relative
to prices of all grades for the whole flue-cured belt, types 11(a)-1h.
Selections for each group (better, medium, or poorer) were limited to grades
that fell in the same third of the range of season average grade prices for
the flue-cured belt as a whole for at least 5 years between 192 and 198.
Grades having prices that fell in the highest third of the season average price
range for types 11(a)-1) are referred to as better grades, those in the middle
third of the range as medium grades, and those in the lowest third as poorer
grades. Of these grades, only those that met this standard for one of the
quality groups in all of the four types of flue-cured tobacco were usede The
selection of United States grades, therefore, is identical among the types of
flue-cured tobacco for each quality group.

This method of grade selection assumes that relative price of a United
States grade during a period of years is a fairly accurate measure of quality
or relative usefulness of the grade compared with other grades. It assumes
that the influence of other factors which also affect leaf prices, such as
the quantities of a grade and of other comparable grades available from the
current crop and the characteristics of the leaf stocks in the hands of manu-
facturers and dealers, balances out during a period of time.

yy U.S. GOVERNMENT PRINTING OFFICE : 1952 O—207987

re a tA

~