TABLE 1.-WHITE PAN BREAD: COMPONENTS OF MARGINS PER 1 LB LOAF, 1972

1 Based on wheat prices received by farmers at country elevator less imputed value of byproducts. Includes the marketing certificate.

2 This averaged 31.8 cents per bushel of wheat (87 percent to flour) based on a 0.4 cent price spread between the farm price and f.o.b. mill cost of wheat in a 1 lb. loaf of bread.

3 Spread between f.o.b. mill sales value of flour (0.6329 lb) used per 1 lb loaf of bread and the f.o.b. mill cost of 0.01445 bushels of wheat less imputed cost of millfeeds.

The price of flour received by the miller deducted from the cost of flour to the baker represents transportation between the 2 points.

Spread between the estimated price received by bakers for bread delivered to store shelf and the estimated cost of flour to the baker. This deviates from marketing spreads received by baker-wholesalers for bread normally reported by ERS, because this includes "other" cost of processing of 2.2 cents for the cost to the baker of nonflour ingredients. Cost compounds are based on National Commission on Food Marketing data, "American Baker's Association" Cost survey data for 380 plants in 1972, and SEC-FTC quarterly financial data.

Source: "Developments in Marketing Spreads for Agricultural Products," Economic Research Service, U.S. Department of Agriculture, March 1974.

TABLE 2.-WHITE PAN BREAD: COMPONENTS OF RETAIL PRICE PER 1 LB LOAF, 1973

1 Based on wheat prices received by farmers at country elevators less imputed value of byproducts. Includes the market-' ing certificate until July 1.

2 Difference between cost of wheat to the miller and farm value.

3 Spread between f.o.b. mill value of flour (0.6329 lb) used per 1 lb loaf of bread and the f.o.b. mill cost of wheat (0.01445 bu) less imputed cost of millfeeds.

Spread between the f.o.b. mill price of flour and the cost of flour to the baker.

5 Spread between the estimated price received by bakers for bread delivered to store shelf and estimated cost of flour to baker. This deviates from the baker-wholesaler spread normally reported by ERS because it includes the cost of nonflour ingredients to the baker of 2.5353 cents. This cost is included in "other" processing costs.

"Transportation from country elevator to flour mill.

Source: "Developments in Marketing Spreads for Agricultural Products in 1973," Economic Research Service, U.S. Department of Agriculture, March 1974.

According to unrevised CED data, the average price for the wheat in a loaf of bread was 2.85 cents in 1972 and 4.15 cents in 1973. The component charts for those years, supposedly based on the same statistics, state that the average price of the wheat in a loaf of bread was 2.8 cents in 1972, and 4.10 cents in 1973. In 1972, this discrepancy was equal to 455 percent of the profit realized in the assembly-of-wheat and 83 percent of the miller's profit. In 1973, this discrepancy was equal to 333 percent of the profit realized in the assembly-of-wheat and 18 percent of the miller's profit.

This discrepancy is apparently a result of rounding off .05 to the next lowest tenth of a cent. Draft notes indicate that the computed farm value was 2.8558 for 1972 and 4.1386 for 1973. As stated in the previous paragraph, the published farm value was 2.8 for 1972.

The cost of wheat f.o.b. the mill minus the farm value of wheat is the determinant for allocating a cost to the assembly-of-wheat spread. This includes transportation and merchandising. It can also include cleaning before it gets to the mill. USDA methodology states that "it is assumed that merchandising spreads covered costs including profits equal to 10 percent of the gross spread. Costs are derived from ERS-513 (Cost of Handling Grain and Controlling Dust in Commercial Elevators, 1971-72 . . . Projections for 1973-74), and related ERS files." The 10 percent figure is in conflict with published component charts, which show the profit of wheat assembly as 2.5 percent of the 1972 assembly-of-wheat spread and 4.7 percent of the 1973 assembly-of-wheat spread. No explanation is given for either the discrepancy or use of the 10 percent.

It is possible that "10 percent" figure represents 10 percent of the merchandising part of the "assembly-of-wheat" spread. Assembly-ofwheat less transportation is .106 cent for 1972 and .12 cent for 1973. Profit is 10.38 percent of .106 cent in 1972 and is 12.5 percent of .12 cent in 1973. These figures are close to the "10 percent" but are not exact. A review of USDA's worksheet yields no more information except for confirmation that it does presuppose that 10 percent of costs are profits.5 There is no clear explanation, however, of how the final results were achieved using such a computation.

Appendix IX is a working table entitled "White Pan Bread: Flour Miller's Spread." Columns 1-4 contain one methodology and columns 5-8 provide an alternative method for computing miller's spread components. The 1973 component chart figures were derived through the second method."

The alternative method starts with the computed mill sales value of flour (3.8107). This is multiplied by .22 percent, the estimate made by the National Commission on Food Marketing (Technical Study No. 5) in 1965 of millers' profits as a percentage of sales value of flour. The result, .0084 cent, is divided as follows: 80 percent or .0067 cent, to processing and 20 percent, or .0017 cent, to wholesaling

4 J. C. Eiland 1973 memo. See Appendix IV, p. 392. 5 See Appendix VIII, p. 401.

• See p. 402.

In method 1, the 1972 USDA published estimates are multiplied by an adjustment factor to make the 1972 estimates more accurate. This factor is based on information, the source of which is apparently, more recent production figures from the 1972 components are multiplied by the BLS indices. Costs are then subtracted from the established spread (.9816 cents) leaving a residual of .3384 cents, labeled profit. This is 5.64 times the profit figure for 1972.

functions on the basis of the same NCFM study. (It should be noted, however, that this approach reduces profits to an astonishing 2 percent of the level arrived at through Method 1.) Under this method the projected profit (.0084 cent) is then subtracted from .6662 cent, the adjusted 1972 spread. The remainder, .6578 cent, is the "new" total cost for 1972 which is allocated to cost components according to the 1972 estimates. To arrive at the new 1972 labor cost, .6578 cent is multiplied by the allocation for processing labor, or 39.1 percent. The "new" labor cost is then multiplied by the Bureau of Labor Statistics (BLS) index to get the published 1973 components. This alternative method calls for taking the 1972 spread and lowering the profit so that costs will increase. These "new" cost figures, having been artificially raised for bookkeeping purposes, are then multiplied by real BLS indices, "raising" the costs sufficiently to cover the 1973 spread. It should be noted profits went up considerably, and they too helped to cover the spread.

Essentially what the 1973 analysis assumes is that the total spread is a given and the millers' profit rate has been constant since 1965. It is then assumed that all component costs have gone up proportionately in one year (all up by about 22 percent), and then small BLS adjustments are made. Indeed, in an earlier memo, one USDA analyst stated:

Cost components are based on costs developed by National Commission on Food Marketing as reported in Technical Study No. 5. Inasmuch as the flour millers' spread has changed very little since that study was made, the cost components have not changed much either.8

There is no evidence available, however, to support this contention. The prose summary of the table for public consumption and finally published in "Marketing and Transportation Situation," November 1974 (MTS-195) states only that "estimated profits for assembly wheat and flour milling were higher in 1973 than in 1972." The methodology memorandum states that miller's 1973 profits went up "three times 1972 levels." In fact, according to the alternative method, they went up 4.6 times.

B. BAKER/WHOLESALER AND RETAIL SPREADS

The next element following the flour miller spread to be covered by the cost analysis is transportation of flour to bakery. This is determined by subtracting the f.o.b. mill price of flour from the cost of flour to the baker. Both these figures are data collected from a sample of flour milling plants. The baker pays the transportation cost; however, since the function is actually carried out by neither the miller nor the baker/wholesaler, it is represented as a separate marketing entity in the cost component charts for 1972 and 1973.

The next major component of the chart is the baker/wholesaler spread. Three fixed elements must be considered:

1. Wholesale price (22.2 cents, as reported by BLS in the Wholesale Price Index);

2. Cost of flour to baker (5.7 cents or the cumulative sum of the cost components of the analysis up to this point; and

8 See J. C. Eiland 1972 memo. See Appendix V, p. 395. J. C. Eiland 1973 memo. See Appendix IV, p. 392.

3. Baker/wholesaler profit [(.5832 cents per loaf as published in the Federal Trade Commission Quarterly Financial Reports (FTC-QFR)].

The baker/wholesaler spread will be 16.5 cents (22.2 cents minus 5.7 cents). The cost components for manufacturing and wholesaling must add to this figure.

The percent breakdown of that 16.5 cents into the various cost components, is derived from three figures:

1. National Commission on Food Marketing cost divisions as reported in Technical Study No. 5;

2. American Bakers' Association (ABA) cost survey data; and 3. The BLS annual indices of change.

ERS does not collect data directly from the baking industry. When asked if they had any industry input on costs, reference was made to the ABA cost survey. Memoranda and footnotes likewise cite the survey. ABA data was used for three purposes in producing the published 1973 figures. These uses emerge from an examination of the worksheet.10

Several of the percentage change indices for both manufacturing and selling are derived from the ABA survey. The 10 percent increase from 1972 to 1973 in labor costs for manufacturing and selling is reflected in the ABA survey. Likewise, the ABA data indicates no change in packaging costs, an assumption that is preserved in the spread data. The business tax change index is also labeled on the worksheet, as derived from ABA data, however, the survey does not list such a cost component. The only figure which shows a change is that of Administrative Expenses (7.3 percent). (In the ABA cost survey instructions, taxes are not mentioned as coming under "Administrative Expenses"; but, apparently the methodology used merely applied the ABA "Administrative Expense" category to the USDA category "Business Tax.") The same 7.3 percent change index is applied to the advertising component even though there is a component in the ABA survey headed, "All Other Selling and Advertising Expenses" (which includes all direct advertising expenses). No explanation is given for use of this figure. In these instances, it appears that the use of the ABA data was quite arbitrary.

Additionally, no change is made in "other" manufacturing costs, on the basis of lack of change indicated by the ABA. However, for "other" selling costs, the BLS index is utilized, perhaps because the "other" selling heading in the ABA data includes advertising.

The ABA survey results are used to determine the percentages of the spread to be allocated to manufacturing and sales costs. It is somehow determined that the ABA data show that 36.2 percent of the costs are in manufacturing and 63.8 percent are in selling. Such a determination required discussions. The ABA data is not broken out in sufficient detail to identify each manufacturing expense and each selling expense. Clear manufacturing costs in ABA data add to 27.7 percent; selling to 59.4 percent. 12.9 percent is accounted for by "Indirect Wages" and "Administrative Expenses." On some unknown basis, 8.5 percent of that portion was assumed to be manufacturing expense and 4.4 percent selling.

10 See Appendix X, p. 403.

On the basis of these percentages 36.2 percent of the 13.4 cents spread (16.5 cents less profit and other ingredient costs) or 4.8474 cents, unrounded, was then allocated to manufacturing, and 63.8 percent or 8.5434 cents, unrounded, was allocated to selling.

FTC-QFR data indicating 2.625 percent profit on the wholesale price is used, thus making overall profit .5832 cents. To allocate this overall profit figure to the manufacturing and selling sectors, the methodology assumes that profit bears the same proportional relationship to the spread as do costs. Thus, it is concluded that 36.2 percent of the overall profit, or .2111 cents is made by manufacturing and 63.8 percent of this profit figure, or .3721 cents, is made by the wholesaling sector. Again, no justification for the assumption is offered. While there may be no other basis for an allocation, some justification would be appropriate.

There is a further general problem with the use of ABA cost survey data. The ABA's survey covers all loaf varieties, brown and serve bread-type rolls, etc., while ERS figures, partially based on ABA data, are supposed to be representative of costs on 1-pound white pan bread. There is no evidence that costs for all loaf varieties are comparable to costs on white pan bread in 1-pound loaves. Some discussion of this assumption is certainly in order. There may very well be substantial differences in cost of production and wholesaling of different bread products. Perhaps it is true that, with limited resources, the ABA data is the best that ERS can do for industry data. Nevertheless, nowhere in the published materials nor in the work sheets or memoranda, appears any qualification, or any admission that the data is not strictly appropriate for the use to which it is put.

The indices which are used to update component percentages of depreciation, rent, repairs, energy, and other selling expenses are derived from BLS. They are not product-specific; they represent general changes in the cost of doing business. The one remaining change index, that which is applied to interest expense in both manufacturing and selling, is computed with the BLS index as one factor. The index is multiplied by the index of change in the price of flour to the baker. This approach was adopted as a result of the dramatic rise in flour prices during the relevant period (i.e., the baker in 1973 was forced to borrow substantial sums of money to maintain the same level of flour in storage. Thus, interest payments were proportionally higher.) A peculiarity which develops, however, is that the same change index is applied to the interest component of selling expenses. Cost of flour can have no possible effect on size of loans taken out by the baker to cover his wholesaling costs.

The final 1973 baker/wholesaler figures are arrived at through four complex steps which need not be described at length here. Let it suffice to say: There was a preliminary conclusion, based on ABA and BLS cost data, which should have been an accurate spread, but instead resulted in a spread far exceeding the known baker/wholesaler share in 1973. The preliminary spread had to be "reduced." This was done by arbitrarily lowering its components (costs and t) based on components share of the "excess spread." The retailing spread, or the difference. price, 22.2 cent, and the retail price, 27.6 ceed or reports) remains. That is 5.4 cent is brok