Up to this point it has appeared as if, _Railroads: Rates and Regulations

the constant and variable expenditures of a railroad, it was the

latter only which grew as the volume of traffic increased. This is

not absolutely, but only relatively true, not only of the so-called

constant operating expenses, but of fixed charges as well.[39]

Everything depends upon the length of time under consideration. Many

expenses follow the fluctuations of business, not evenly but by

jerks. Up to the full limit of utilization of the existing plant,

each increment of traffic seems to necessitate but a very small

increase in the so-called variable expenses, with hardly any change

at all in the constant ones. A branch road can haul more and more

tons of freight with a given outfit of cars and locomotives by merely

increasing slightly its outlay for fuel, train service, wages and

supplies. But after a certain point more rolling stock must be provided

to accommodate the growing business. As each of these additions to

property occur, they contribute new quotas to the fixed charges and

to the so-called constant expenses of operation, such as maintenance

of roadway and the like. Nor can these new expenses be allocated to

the new business alone. The moment the old traffic has outgrown the

existing plant, the new expenditure becomes chargeable to all the

business alike. The new outgo must be distributed evenly over the

entire volume of traffic thereafter handled. Each ton, both of old and

of new traffic, beyond the haulage capacity of the locomotives then

in service, is equally responsible for the expense of new equipment

purchased. Although the old business could have been handled without

a million dollars spent for double-tracking or terminal enlargement,

this addition to the expense of maintenance of way or to the fixed

charges is equally attributable to every ton of traffic hauled.

A concrete example may aid in making this important principle clear.

The new through-freight trunk line built by the Pennsylvania Railroad

since 1900, paralleling its old four-track one, represents both in the

cost of maintenance and capital charges, a sudden jump in the expense

of transporting each ton of freight on _both_ lines, until such time

as the new business grows to a point where it can support the new line

by itself alone. The relation between increasing returns and density

of traffic is well illustrated in this instance. With six tracks in

operation nearly all the way from Pittsburg to Philadelphia, the four

old tracks are sometimes almost fully utilized for passengers and fast

freight. The extraordinary density of traffic appears in the statement

that this road in 1911 on 3534 miles of track handled one-third more

ton miles than the Union Pacific--by far the most worked of all the

western lines--handled on 13,674 miles of track. The two new low-grade

Pennsylvania freight tracks are used only for slow traffic; largely

coal and westbound steel empties. Not-withstanding the extraordinary

density of traffic on this extra two track line, it probably does

not meet the fixed charges on cost of construction of the line. Yet

the new double track was absolutely necessary, regardless of its

profitableness, in order to relieve congestion on the old four tracks.

In other words, the demands of the service forced an expenditure which

in and of itself was not financially self-supporting. But the profit

from the old lines would be sufficiently enhanced to take care of

the whole. The bearing of such cases upon the capital needs of the

future is obvious. A resolutely conservative policy of finance becomes

imperative under such circumstances.

In much the same way, the general condition of congestion reached

in 1903-'05 on the eastern trunk lines and in the West and South in

1906-'07, manifested mainly in the need for more tracks and terminals,

represented the permanent outgrowth of the old plant; and necessitated

a readjustment of capital expenses for the purpose of enlargement.

Viewed in a large way over a term of years, nearly every expenditure,

even the fixed charges which appear constant or independent of the

volume of business, thus become in reality imbued with more or less

variability.

The preceding considerations hold good not alone of increased

facilities, but of their curtailment as well. This point is often

neglected in respect of capital outlay, which once made cannot be

recalled. Rotting of ties we have held to be a constant expense

of operation. It goes on steadily, whether traffic conditions be

good or bad. But, on the other hand, those ties, if they be under

a third or fourth track, would never have been laid had not there

been a promise of business sufficient to render the added investment

profitable. As Lorenz observes, "the question is not, What expenditures

would disappear if a certain proportion of the traffic should be

discontinued? but What expenditure would not now be incurred if that

traffic had never been called forth?" Viewed in this way, even the

necessary replacement of ties under a (temporarily) little used extra

track, is an expense determined at some time, even if not always, by

the volume of the business. In the long run, therefore, the percentage

of total cost which we may assign to an increase in the volume of

traffic, is higher than appears from a cross-section of expenses,

taken, as was at first had, in a given year. Lorenz has illustrated

this steady expansion of all groups of expenditure in relation to

expansion of traffic by the following table, in which the actual

figures for each year [brought down to date] are replaced by an index

number based upon 100 for the year 1895. It would have been highly

suggestive to continue all of this data alike to the present time; but,

as noted on the table, certain items have been so modified by changes

in accounting practice, that this could not be done.

==================================================================

|Gross earnings from operation

| |Ton miles

| | |Passenger miles

| | | |Total operating expenses

-----+-----+-------+-------+-----+-------+-------+-------+--------

1895 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | 100

1896 | 107 | 111.8 | 107 | 106 | 111.2 | 117.9 | 103.1 | 99.4

1897 | 104 | 111.6 | 100.5 | 103 | 108.5 | 106.4 | 99.8 | 98.4

1898 | 116 | 133.8 | 109.7 | 113 | 120.4 | 124.9 | 109.1 | 101.1

1899 | 122 | 145.1 | 119.7 | 118 | 126.8 | 134.6 | 115.6 | 110.0

1900 | 138 | 166.1 | 131.5 | 132 | 150.4 | 164.2 | 126.9 | 112.7

1901 | 147 | 172.5 | 142.3 | 142 | 164.6 | 173.2 | 135.5 | 121.8

1902 | 160 | 184.5 | 161.5 | 154 | 185.2 | 200.2 | 151.7 | 131.4

1903 | 176 | 203.2 | 171.6 | 173 | 198.6 | 225.6 | 174.7 | 142.1

1904 | 184 | 204 | 179.8 | 184 | 194.8 | 250.7 | 188.6 | 153.5

1905 | 193 | 219 | 195 | 191 | 191 | 253 | 179 | 154

1906 | 216 | 254 | 206.5 | 212 | 216 | 288.8 | 194 | 166

1907 | 240 | 277 | 227 | 241 | --[40]| --[40]| --[40]| --[40]

1908 | 222 | 256 | 238 | 230 | -- | -- | -- | --

1909 | 224 | 256 | 238 | 220 | -- | -- | -- | --

1910 | 256 | 300 | 265 | 251 | -- | -- | -- | --

==================================================================

According to this showing, maintenance of equipment, which we held in

our analysis to be about one-half a constant expense and independent

of traffic, especially after 1900, appears to have actually outrun

the expansion of ton-mileage and passenger business. How largely

this is due to actual purchases for the sake of future growth is

not determinable. And maintenance of way outlay--one of our largely

constant expenses--has increased, in fact, more rapidly than conducting

transportation, which we held to be mainly variable. But these

figures are confused by the failure to differentiate in the accounts,

mere maintenance from actual improvements and additions to plant.

Expenditures for these latter purposes, charged to operating expenses

rather than to capital account, have been so enormous during these

years of prosperity that they confuse the true facts utterly. It is

to be hoped that now with the revised statistics since 1906, which

will permit a clearer definition of these expenditures in detail,

an analysis covering a series of years will bring out the real

relationships. Equally important is the fact that these years have been

characterized by rapid and extensive rises, both of prices and wages.

Had our table covered a longer series of years the results would have

been more clear. Until such an analysis be made, it will suffice for

our purpose, viz., the analysis of the principles of railroad rate

making, that we adhere to our first general conclusion, namely--that

of the total expenditures of a railroad _at any given time_ about

two-thirds of them are constant, while only one-third vary with the

ups and downs of the volume of traffic. Comprehending in survey a long

period of years, it might happen, as Acworth concludes, that nearly

one-half of the total expenditures were entirely fixed in character,

leaving the other half as dependent upon the amount of transportation

effected.

The manner in which heavy capital outlay for maintenance accompanies

as well as partly accounts for a decline in the cost of conducting

transportation on American roads, is graphically shown by the diagram

on the next page.[41] During ten years a steady decline in direct

operating costs has accompanied an equally marked upward tendency in

expense of maintenance. The bearing of this on the problem of rate

advances in future is direct. Profitableness results from two separate

sources; economical operation such as longer trains and better loading,

and also from far heavier capital investment in plant, by which such

operation is rendered possible.[42] Both alike, however, attend upon

increased volume of business. Heavy capital investment may lessen

immediate maintenance charges,--lower grades and straighter alignment

naturally wearing less; but, on the other hand, the burden of interest

and other fixed expenses steadily grows. How will they stand toward

one another by 1925 on the eastern trunk lines? Will growth of business

bring lower rates or not? A fine field for further analysis is as yet

unworked.

[Illustration: _RATIO OF MAINTENANCE OF PROPERTY AND CONDUCTING

TRANSPORTATION TO TOTAL OPERATING EXPENSE._]

One final relation between operating and fixed expenses is left for

consideration. It is so well put by J. Shirley Eaton in an unpublished

paper, that it can best be stated in his own words:

"It is impossible to have an absolute and universal line of

demarcation between the direct and the fixed expense, that

shall be the same on all roads. One road chooses to reduce

a grade and thereby increase the capital account in order

to save in the current expense of a helper at a hill or the

lost margin of efficiency of the loaded train on the level.

The relation between a current expense and the annual charge

of the capitalized cost on a fixed plant that performed the

same service, was well illustrated in a case arbitrated by

Mr. Blanchard in New Orleans. One road which did not have

access to the heart of the city undertook to compensate its

disadvantage by trucking to and from its depot. The hire of

a public truckman to perform the service for its patrons was

very soon commuted to the practice of paying the amount of

the truck expense to the consignee by deducting it from the

freight bill rendered, the consignee or shipper performing the

service. This, known as 'drayage equalization,' was claimed by

competitors to be in the nature of a rebate to secure business.

The arbitrator decided that the first roads had elected to buy

their right of way into the heart of the city; and the road

that had not built into the city elected to pay the expense of

the same service in the shape of a current drayage bill instead

of in the shape of interest on money invested in right of way.

Therefore he decided there was no cause for complaint."[43]

Railroad expenditures, as Taussig clearly pointed out a number of years

ago,[44] afford a prime illustration of the production of several

commodities by a single great plant simultaneously at joint and

indistinguishable cost. The classic economists illustrated this law by

the joint production of wool and mutton and of gas and coke. In both of

these instances neither commodity could conceivably be produced alone.

Nor was either one, so to speak, a by-product of the other. So nearly

of equal importance are the two, in fact, that the cost of production

for each may approximately be determined by dividing the total cost

according to the relative worths of the two or more products. The

law of joint cost with reference to the production of transportation

is somewhat different. Compare, for instance, the carriage by a

railroad of thousands of passengers and different commodities in every

direction, under varying conditions, singly or by wholesale, slowly or

by express, over a given set of rails every day; with the operation

of a great refinery producing simultaneously kerosene, gasoline,

lubricating oils and greases as well as various odd chemicals. Both

are examples of production at joint cost, but with various important

contrasts. In the refinery all the costs are joint. All the processes

are interlocked. Every increase in the output of kerosene produces

_pari passu_ an increase of the other commodities. On the railroad not

all, but only a part of the costs are joint, in such manner as has

been shown. For, from the joint portion of its plant--roadway rails

and locomotives--the railroad may produce transportation of different

sorts quite independently. It may choose to especially cultivate its

passenger traffic, or its cotton or coal business. After a certain

point of congestion is reached, the various sorts of traffic on the

railroad may even become actually competitive with one another so far

as the joint use of the plant is concerned. It is plain that this

could never happen in the refinery. The use of more stills for making

kerosene would automatically produce more by-products of every sort.

But on a railroad it might well happen that the coal and passenger

business might come to interfere with one another. A choice of emphasis

as between fast refrigerator beef or fruit traffic, and limited express

service, may have to be made on a long single track line. Nevertheless,

in spite of these peculiarities of transportation, the general law

of joint costs holds good, in that it is a demand for each service

rather than its cost which finally determines the chargeable rate.[45]

This must be so, because of the fact that the cost of each shipment

is so largely joint and indeterminate, and that a large part of the

entire plant is indistinguishably devoted to the general production of

transportation without reference to particular units of business. One

concrete example may serve to illustrate this point.

For years attempts have been unsuccessfully made by accountants to

effect the primary separation between expenses of passenger and freight

business,[46] in order to determine the cost of transportation per

unit in each case. Some companies like the Louisville & Nashville

and the Burlington system, still divide up the two, usually on the

basis of the engine mileage for each class of traffic. This may be

serviceable enough for comparisons of costs from year to year in the

same company, but it has no general value and it may, moreover, become

highly misleading. The most absurd conclusions may result. Thus at one

time it appeared from such data, compiled by the Interstate Commerce

Commission, that the New York Central, with five times the density of

traffic of the Illinois Central, was actually conducting its freight

business at a much higher cost per ton mile. Such inconsistencies

induced the Interstate Commerce Commission in 1894 to abandon the

attempt at any such primary separation of accounts.[47] It has since

been reattempted, in special cases, as by the Wisconsin Railroad

Commission in its notable "Two-cent Fare" decision in 1907, the

division being made according to a number of different criteria.[48]

But it is plain that a very large proportion--probably over half--of

the expenditures for freight and passenger business are entirely joint,

however distinct the revenues from each service may be. We have seen

that approximately two-thirds of the outgo is incurred on behalf of

the property as a whole. Certain expenses, to be sure, such as train

wages, coal consumption and the maintenance of rolling stock, are

readily divisible; but with respect to the maintenance of way and

structures--about forty per cent. of the total outgo--all guides fail.

Even in respect of the cost of rails, due to wear and tear of train

movement, we are quite at sea in the allocation of expenses. Freight

trains may indeed be four times as heavy as passenger trains; but, on

the other hand, they move at far slower speeds. And then, finally, how

about the large item of capital cost, the proportion of outgo for

fixed charges? This equals about twenty-seven per cent. of the total

expenditures for the United States as a whole. We may, of course,

divide these expenses arbitrarily on the basis of the relative gross

revenue from freight and passenger business respectively. And yet how

absurd it would be to attempt to allocate an expense of a million

dollars for the abolition of grade crossings in this way. As between

the New Haven road, with passenger and freight revenues about equal,

and a western road with only one-tenth of its income derived from

passengers, the apparent cost of freight business on the eastern road

would be absurdly reduced by any such process. The facts are plain.

So many expenditures are incurred indiscriminately on behalf of the

service as a whole--being an indispensable condition for operation

of the property at all--that no logical distinction of expense even

as between passenger and freight traffic is possible. This being so,

how futile it is to expect to be able to set off the expenses due to

any particular portion either of freight or passenger service, and

especially to any individual shipment. It may oftentimes be possible

to determine the _extra_ cost due to individual shipments. This, of

course, mainly applies to what are called movement expenses. Thus the

haulage cost of a 2,000-ton grain train from Chicago to New York has

been estimated at $520. But how small a part this is of the total

cost, the preceding analysis must have made clear. In the Texas Cattle

Raisers' case, detailed analysis of the extra cost for the traffic

in cattle was presented.[49] The starting point in this attempt was

necessarily an allocation of freight and passenger expenditures, which,

if defective, would vitiate the entire subsequent calculation as to

costs. In this instance, it was the judgment of the Interstate Commerce

Commission in its final decision in 1908, that no such separation of

expenditures was possible as a basis for the determination of cost of

service.

FOOTNOTES:

[29] _Quarterly Journal of Economics_, XXII, 1908, p. 364 _et. seq._

[30] U. S. Statistics of Railways, 1908, p. 165 (and annually

thereafter), gives an outline of these expense accounts for all

railways over five hundred miles long.

[31] Treated in vol. II, chap. XV. Begins in U. S. Statistics of

Railways, 1909, p. 76.

[32] Changes in accounting rules in 1907 prevent its continuation

to date; but the data for 1909 under the new system are reproduced

alongside.

[33] U. S. Statistics of Railways, 1908, p. 165, and annually

thereafter gives data for all large roads.

[34] The sharp decline in traffic in 1911, especially after the

suspended advance of rates, as affecting maintenance expenditures per

mile of road, is shown as follows:

===================================

| 1911 | 1910

-------------------+-------+-------

Baltimore & Ohio | $5931 | $6336

Union Pacific | 3296 | 3363

Great Northern | 2375 | 2653

New York Central | 8681 | 8087

Northern Pacific | 2451 | 3413

Pennsylvania | 9088 | 9792

Multiplying these differences into thousands of miles of line shows the

great economy resulting.

[35] _Cf._ pp. 259 and 422, _infra_.

[36] The provision of plant and equipment to carry the "peak of the

load" is often a serious handicap.

[37] For an instance of detailed analysis of cost, the general

investigation of soft coal rates to the lakes in 1912 is highly

suggestive. Two-thirds of revenue went for operation and maintenance,

one-third for return upon plant. This was the first attempt to justify

an advance in rates for a large volume of traffic on the ground that it

did not contribute its proportionate share of earnings. 22 I.C.C. Rep.,

604.

[38] From Railroad Operating Costs; by Suffern & Co., New York, 1911.

[39] Lorenz in _Quarterly Journal of Economics_, XXI, pp. 283-292, is

suggestive.

[40] Change of accounting methods vitiates further comparisons of

operating costs after 1907.

[41] From Railroad Operating Costs, by Suffern & Co., New York, 1911.

[42] _Cf. Yale Review_, 1910, pp. 268--288; with reference to the

rate advances of that year.

[43] _Cf._ the Free Cartage case, 167 U.S., 633.

[44] _Quarterly Journal of Economics_, V, 1891, pp. 438-465.

[45] Two important qualifications of this law, however, are set forth

at p. 265, _infra_.

[46] _Cf._ our Railway Problems, rev. ed., _circa_ pp. 684, 706.

[47] The first successful attempt, as to soft coal rates to the lakes,

is in 22 I.C.C. Rep., 613. Cf. 13 _Idem_, 423.

[48] Wisconsin Railroad Commission Report, 1907, p. 101. Compare also

Woodlock, p. 91; U. S. Statistics of Railways, 1894, p. 70; _Yale

Review_, 1908, p. 382; and Record, Cincinnati Freight Bureau Case, II,

p. 941.

[49] 13 I.C.C. Rep., 423. Compare 9 _Idem_, 423; and _Yale Review_,

1908, p. 287.

CHAPTER III

THE THEORY OF RAILROAD RATES (_Cont'd_)

The law of increasing returns, 71.--Applied to

declining traffic, 73.--Illustrated by the panic of

1907, 75.--Peculiarly intensified on railroads, 76.

Growth of mileage and traffic in the United States since

1889, 77.--Increase of earnings, 79.--Operating expenses,

gross and net income, 80.--Comparison with earlier

decades, 85.--Density of traffic, 86.--Increase of train

loads, 88.--Limitations upon their economy, 92.--Heavier

rails, 93.--Larger locomotives, 94.--Bigger cars, 95.--Net

result of improvements upon efficiency and earning power,

97.

The law of increasing returns due to financial rather than

operating factors, 99.

A railroad theoretically presents a clear example of an industry

subject to the law of increasing returns--that is to say, an industry

in which the cost of operation grows less rapidly than the volume

of business done. Each ton of freight added to the existing traffic

costs relatively less to haul. From this it follows, obviously, that

the net returns increase more than proportionately with the expansion

of traffic. This may be demonstrated by a simple calculation. It has

already been shown that only about two-thirds of the total expenditures

of a railroad are applied to operation, the remaining third being

devoted to capital account. Moreover, of these two-thirds of the total

applied to operating outlay, only about one-half responds to any change

in the tonnage, the other half being constant up to a certain point.

Otherwise expressed, an increase of one per cent, in traffic and,

therefore, of revenue, produces an increase in expense of only one-half

of two-thirds of one per cent.[50] Two-thirds of the entire increment

of revenue goes to profit. Carry this increase further and the effect

is more striking. Suppose traffic to grow tenfold. The former outlay

being $100 for a given volume of business, would be divided according

to our rule as follows: one-third for fixed charges, one-third for

constant operating outlay and one-third for variable expenses. With ten

times as much traffic, only the last group of outgoes will expand. One

thousand dollars revenue would therefore become available under the new

conditions, to pay the same fixed charges as well as constant operating

costs. The total outgo would thus become $33 plus $33 plus $330, or

$396 in all. Almost two-thirds of the increment of revenue still

remains as profit. It might well happen that such an expansion could

not ensue without large increases in the capital and plant, as has

already been noted; but up to that point this calculation would hold

good. The following statement varying but slightly from our foregoing

assumptions, illustrates the principle.[51] Let the distribution of

expenditures for given conditions, producing $100 of revenue, be these,

viz.:

Operating expenses $ 67

Fixed charges $ 28

----

$ 95

Profits for dividends $ 5

----

$100

Now assume an increase of ten per cent. in the traffic and consequently

in the revenue; but assume also that the average _extra_ cost per

unit, of the new business, is only forty per cent. as much as for the

preëxisting tonnage. Were the added cost of each ton mile as great as

before, the operating expenses would rise by the full ten per cent. of

$67. But on Webb's assumption, they will rise by only forty per cent.

of ten per cent. The new account would then stand thus:

Operating expenses ($67 plus forty per cent. of

ten per cent. of $67) $ 69.68

Fixed charges as before $ 28.00

-------

$ 97.68

Income, increased by ten per cent. $110.00

-------

Balance for profit or dividends $ 12.32

By an increase of ten per cent. in tonnage, balance for dividends has

more than doubled.

In this connection it will be noted that a constant rate of return per

unit of business newly acquired has been assumed. Attempts were made

on behalf of the railroads, during the long period of decline of ton

mile revenue down to 1900, by Newcomb and others, to show that this is

an unreasonable assumption; in that increased traffic is presumably

to be had only by a progressive lowering of the rates charged. This

contention has been effectively demolished by the steady and remarkable

growth of traffic since 1900, even in the face of a substantial rise

of rates all along the line. A necessary corollary to our proposition,

beside that of the maintenance of a constant scale of charges, is, of

course, also of the continuance of a given grade of service and of

costs of operation. If more luxuriously appointed passenger trains or

quicker freight service have to be given in order to produce the growth

of business, the added costs of operation must, naturally, be taken

into consideration. If widespread rise of wages follows an increase in

the general cost of living, that too is an entirely extraneous factor.

But with a given grade of service, constant rates and steady wage

scales, there can be no question, up to the point of full utilization

of the existing plant, that the operation of railroads affords clear

demonstration of the law of increasing returns.

The obverse side of the law of increasing returns is also of great

importance. For the same reason that when traffic increases, only a

portion of the expenses are affected, it follows that, when business

declines, only a part of the costs can be lopped off. In other words,

a reduction in the volume of traffic does not in itself alone lead to

a corresponding reduction in the operating expenses. Of course, many

of these latter may, as we have seen, be temporarily postponed, as

they were in 1893-1897, especially in the group of maintenance-of-way

expenses. In such an event they must ultimately be made good by

extraordinary outlay at some later time. But, unless they be thus

postponed and unless the rates charged for service be reduced in order

to stimulate traffic, it is inevitable that the margin of profit will

drop as rapidly as it tends to rise with increased volume of business.

This may be illustrated by the following computation.[52] Assume the

total revenue from a given business to be $100, and assume it to be

distributed as before, viz.:

Operating expenses $ 67

Fixed charges $ 28

----

$ 95

Leaving profit $ 5

----

Total $100

A positive decline of ten per cent, in the tonnage, if the cost for

operation per unit of the portion lost was the same as the rest, would

obviously reduce the operating expenses also by ten per cent. Let it

next be assumed, as was done previously, that the average extra cost

per unit of the latest increment of business was only forty per cent.

as much as for the remainder of the tonnage. How closely this will

approximate the facts in any particular instance will depend upon the

density of traffic attained in relation to the capacity of the existing

plant. If the addition of the last ten per cent. of business did not

increase the large proportion of fixed expenses at all, and only added

forty per cent. per unit more to the variable expenses; per contra, the

loss of it would merely reduce the variable expenses and still leave

the constant outlay the same. On this assumption, by the loss of ten

per cent. of business the total amount of operating expenses under the

new conditions would be lessened, not by ten per cent. of $67, but

by only forty per cent. of ten per cent. of $67. The income would,

however, decline by the full amount of ten per cent. The account, after

a loss of ten per cent. of business, would then stand somewhat as

follows:

Operating expenses ($67 less forty per cent. of

ten per cent. of $67) $64.32

Fixed charges, as before $28.00

------

$92.32

Income, reduced by ten per cent. $90.00

------

Leaving a deficit of $2.32

Or, in other words, a decline of ten per cent. in tonnage has

transmuted a five per cent. dividend condition into one involving an

actual deficit nearly half as great as the former profit. The sudden

reversal from apparent prosperity to very real distress, such as

occurred during the fall of 1907, is thus explained. Its suddenness

may be shown by the following table of monthly gross and net earnings,

promulgated by the Interstate Commerce Commission.[53] The acute panic

occurred during October, but its effect was not apparent until the

following month. The total mileage included is shown by the first

column:

===============================================

| | | Earnings

| | | --per mile--

| | |-------------------

| | Mileage | Gross | Net

-----+-----------+---------+---------+---------

1907 | July | 223,900 | $1,022 | $304

1907 | August | 224,100 | $1,079 | $345

1907 | September | 224,300 | $1,045 | $314

1907 | October | 224,700 | $1,116 | $337

1907 | November | 224,800 | $981 | $261

1907 | December | 224,400 | $861 | $197

1908 | January | 198,700 | $746 | $148

===============================================

This table shows that whereas under full prosperity, up to and including

the month of October, the net revenue was about thirty per cent. of

gross; after the sharp decline in traffic, it dropped in November to

twenty-six per cent., and progressively thereafter to twenty per cent.

in January. In other words, a decline of about one-fourth in the gross

revenue within four months, entailed a loss of over fifty per cent.

in net earnings. Higher operating expenses in the winter may have

exaggerated this tendency, but, on the other hand, drastic economies

were put into effect, which would more than offset the difference.

The urgent need of at once meeting any loss of business by prompt

reduction of operating expenses is apparent. But there is comfort to

be found at this point in the fact that each one per cent. saved in

operation at any given time, results in saving two per cent. for the

net earnings. According to our estimates, and as a rule practically,

operating expenses equal about two-thirds of gross revenue, leaving

one-third to meet charges and pay dividends. Every reduction from

this two-thirds of gross revenue, therefore, transferred to the

balance, increases the latter proportionately twice as much. This

fact in turn explains the urgent pressure always brought to bear at

such times to effect economies all along the line. These are too

often indiscriminately made.[54] Such paring down of expenses should

always be made with an eye to their ultimate effect upon the operating

efficiency of the property in the long run. To postpone much-needed

repairs of equipment during a period of depression, like that of

1907-1908, when repair shop costs are at a low ebb, only to hamper

operations and to effect repairs under pressure when business revives,

is an instance of such wasteful economy.

The qualification of the law of increasing returns as applied to

railroads, arising from the distinction between long and short term

production of its commodity--transportation--as above described, is

of course by no means confined to carriers alone. It holds good of

a factory or mercantile establishment as well. But in the case of

railways, it is emphasized by the abruptness with which the condition

of congestion of plant arises. The limit of full working capacity

in a factory is elastic, by reason of the fact that under the "peak

of load"--in busy seasons--it may prolong operations beyond the

daylight hours or, at worst, work all night by double shifts. But a

railroad, customarily working by night as well as by day and thus

distributing its operations over the entire twenty-four hours, enjoys

no such expansible limits upon utilization of its plant. When such

full utilization is attained, the end comes suddenly. No postponement

to a more favorable time for raising funds for better terminals or

four tracking the main line is possible; nor does its character as a

public servant permit a railroad to curtail service. The dead wall

of congestion cannot be gotten around by either path. A crisis is

presented, calling for the most heroic measures. This, of course, still

further emphasizes the need for a long look ahead into the future with

respect to railroad finance; not for the management alone, but for the

government as well, charged as it is at present with control over rates

for service.

* * * * *

The application of the law of increasing returns to railroads in actual

practice is beset with difficulties. In order to make these clear, it

will be necessary first to describe the phenomenal development of this

country which has taken place during the last two decades.

The freight service of the railroads of the United States, measured by

weight, in 1910, amounted to 1,026,000,000 tons. Only since 1899 when

the corresponding figure given by the Interstate Commerce Commission

was 501,000,000 tons, have accurate data been obtainable. This would

indicate a growth in ten years of about one hundred per cent. But

this figure takes no account of the distance each ton of freight

travels. This factor is included in what is known as ton mileage--that

is to say, the equivalent of the number of tons of freight carried

one mile. Obviously, so far as the amount of service rendered is

concerned, one ton carried a hundred miles is the equivalent of one

hundred tons transported one mile. Every carrier totalizes in this

way each ton of freight movement by multiplying it into the distance

transported. For the United States as a whole, this ton mileage in

1910 was 255,016,000,000--that is to say, the service rendered would

be represented by the carriage of that number of tons one mile.

The appended diagram shows the phenomenal rapidity with which this

transportation service has grown since 1899. The scale on the left hand

side of the chart serves this purpose. The right hand scale indicates

the miles of line in operation.

[Illustration: Relative Growth of Mileage and Traffic]

The rapid growth up to 1893 was suddenly interrupted by panic and

subsequent industrial depression lasting for about four years.

Recovery began in 1897, since which time the freight movement has

increased by leaps and bounds from about 95,000,000,000 ton miles

to 255,016,000,000 ton miles in 1910. It is obvious that the growth

of transportation in any country is bound to be more rapid than the

increase either in population or in wealth. It appears, indeed, almost

as if the volume of transportation in the United States increased

more nearly as the square of population than in direct proportion.

It has been estimated that we forward two and a half times as much

freight _per capita_ as some of the leading European countries like

France. Our domestic population from 1890 to 1910 increased about

fifty per cent. The railroad mileage grew at about the same rate. Yet

the freight service surpassed this rate of growth more than six times

over; and the passenger service augmented nearly as much. Both alike in

1910 were practically three times as great in volume as twenty years

before. The diagram on page 78 is intended to illustrate the relative

rapidity of this development. While population and mileage increased

about one half, the railroads in 1910 hauled the equivalent of three

times the volume of freight traffic handled in 1890. At the beginning

of this period, the railroads had to seek the freight. Now it appears

that traffic normally will seek the railroads. At times, even, as in

1906-1907, the railroads have actually sought to escape the flood of

business presented.

The magnitude and importance of the growth of tonnage, as above

described, is revealed by the rapid increase in railroad earnings. The

course of these is shown by the succeeding chart on page 82. Gross

revenues of American railroads in 1889 were about one billion dollars.

In 1910 they amounted to $2,750,000,000. Thus it appears that gross

earnings almost equalled three times the amount of twenty years ago.

The net income available for dividends has grown even faster. The

increase was, roughly speaking, about five fold; namely, from 101

millions in 1889 to 515 millions in 1910. Nearly three and one-half

times as much money went annually to the owners of railroad securities

as dividends and interest, besides leaving surplus earnings for 1910

of about 222 millions available for improvements and surplus. But the

limit of utilization seems to have been about reached on many roads

in 1906; and an era of extensive new capital outlay to increase the

existing plants and facilities ensued. Indications are not lacking to

show that at the height of activity before the industrial collapse of

1907-1908, such a point of saturation had been reached, especially in

trunk line territory and on the northern transcontinental lines.[55]

On the Northern Pacific, for instance, the ton mileage increased from

2.2 billions to 5.2 billions between 1900 and 1906. The Northwest

was suddenly confronted at that time with the new issue of enlarging

facilities, which had been slowly becoming apparent elsewhere in the

country during the preceding decade. Grain actually rotted on the

ground, and an acute coal famine occurred, because of sheer inability

of the roads to care for the new traffic. Changes in methods of

business also somewhat exaggerated this strain upon the carriers.

Merchants now expect quick delivery to order. They object to stocking

up months ahead, even when conditions are auspicious; therefore,

business, when especially stimulated, comes with an irresistible

rush. All these causes, coupled with undiscriminating attempts by

inadequately bedded roads to imitate the methods of progressive ones by

prematurely increasing their train loads, led to a practical breakdown

of the transportation business of the country in the autumn of 1906. To

the student of transportation, this congestion denoted the attainment

of a point of saturation for the then-existing physical plant. The

analogy to the case of the Pennsylvania Railroad, previously described,

is obvious. Such a predicament is bound to arise in the development of

any carrier in a rapidly growing country. Its fiscal significance will

appear in due time.

A comparison of the growth of business and of operating expenses for

the entire railroad system of the United States over a series of years

is given in the following table. The results are expressed by means of

index numbers based upon the year 1880, taken as 100.[56]

RELATIVE INCREASE IN TRAFFIC ITEMS, OPERATING EXPENSES AND

REVENUE FROM 1880 TO 1906, INCLUSIVE

======================================================================

| --Average, from and inclusive--

Items | | 1881 | 1886 | 1891 | 1896 | 1904

| 1880| to | to | to | to | to

| | 1885 | 1890 | 1895 | 1898 | 1906

---------------------+-----+--------+--------+--------+--------+------

Ton miles of freight | 100 | 134.36 | 203.23 | 264.90 | 313.81 | 595.0

passengers. | 100 | 138.12 | 189.46 | 233.15 | 224.65 | 412.0

Operating expenses | 100 | 132.75 | 174.39 | 215.30 | 221.42 | 394.0

operation | 100 | -- | -- | 183.0 | 190.0 | 346.0

======================================================================

From this table it appears that between 1880 and 1906 the ton mileage

of freight increased about six fold, and the passenger business more

than four fold. Operating expenses, on the other hand, were in 1906

less than four times as great as in 1880. Increasing returns are quite

evident. The period from 1880 down to 1896-1898, before the recent

general increases in prices and wages took place, shows this even more

strikingly. In order to transport more than three times as much freight

and two and one-quarter times as many passengers, it required a direct

outlay for operation of little more than twice as much money.[57] On

the other hand, owing to the rapid rise of all operating costs since

1898, a comparison of expenditures confined to the last ten years by

themselves, affords an apparent contradiction. The results for this

period have already been given, classified in greater detail. And

yet, despite this disturbing factor and the one earlier mentioned

that these later operating expenses have been heavily loaded with

improvement expenditures, it appears by comparison of 1895 with 1905,

that passenger business has more than doubled, and freight business is

two and a half times as great, while operating expenses in 1905 were

not much over twice their amount ten years before.

[Illustration: EARNINGS & EXPENSES]

A comparison of the movement of gross earnings with operating expenses

introduces still another disturbing factor, namely, the changes from

year to year in the level of freight rates as well as in the character

of the traffic handled. The effect of fluctuating costs of production

of transportation having just been considered, we may now turn to the

fiscal returns as affected by the price obtainable for the service

given. Any long-time comparison of results reflects the influence of

the steady decline of freight rates during the generation prior to

1900. Thus comparing 1880 with 1898, as shown by the preceding table,

operating expenses grew in the ratio of 100 to 221, while gross income

grew from 100 to only 190. Three fold the freight business produced

less than twice the revenue. Pushing the comparison later, down to

1906, operating expenses grew after 1880 from 100 to 394, while gross

income rose to only 346. This reflects the influence during the last

few years of the rapid rise in prices and wages.

* * * * *

According to the opposite diagram, comparing 1890 with 1910, both

operating expenses and gross income from operation seem to have

moved together; the curve of gross revenue rising proportionately

only a little faster than that for operating expenses. The latter

have risen from a general figure of about $800,000,000 before the

depression of 1893-1897, to $1,822,000,000 in 1910; the former from

about $1,200,000,000 to over $2,750,000,000. Both alike somewhat more

than doubled, therefore, in twenty years. At times, especially during

the rapid revival of business after 1897, before rising prices began

to affect costs of operation, extraordinary increases in earnings

appeared, outstripping the growth of expenditures. Comparing the year

1899 with 1895 we find that the gross earnings of the railroads of

the United States increased by twenty-two per cent. This involved

an increased expense of operation, however, of only eighteen per

cent. Similar comparison year by year, there having been an enormous

expansion of business, shows an increase in gross earnings somewhat

more rapid than the growth of operating expenses. This differential

advantage has progressively lessened since 1902, and especially since

the let-up in 1907. The official returns for 1911 with the marked

decline in gross, show an even more distinct drop in net earnings.

Whether the need of an increase of rates commensurate with the

augmented operating costs is imperative, can only be ascertained after

a return to more normal business conditions.

These relationships would be the more striking could we exclude the

enormous expenditures for betterments which have been charged to

operating expenses during these years. Comparisons of net earnings

are vitiated by uncertainty upon this point. Working over these

results by comparison per mile of line, it appears that the rate of

increase in earnings per mile of line for five years prior to 1900,

was approximately double the rate of increase of operating expenses

per mile of line. The greatly lessened cost of performing additional

business becomes at once apparent. But these latter conclusions, as has

been said, cover only a brief period of time. Judging by the results

over many years, it appears that changes both in the level of freight

rates and of wages and prices have operated to leave the railroads not

much better off than they were some time ago. The only thing which has

saved them whole in the face of rising prices and wages since 1900,

and especially since 1907, has been the rise of freight rates and the

enforced improvements in operation. With the methods of transportation,

such as size of cars and locomotives and train loads, as they were a

decade ago, very real distress would be more widely apparent than it

is. On the whole, the public seems to have shared in the benefits of

these improvements to a considerable degree. This statement, however

true for the entire railroad system of the country as a whole, does not

by any means represent the facts for any single system. Moreover, it

is not by any means clear how fully the railroad system of the country

has been enlarged and improved out of surplus earnings. There is reason

to think that foundations in some cases--the Pennsylvania road, for

example--have been laid during these prosperous years, for largely

increased tonnage in the immediate future without a corresponding

growth of expenses chargeable to plant; in other words, that the

transition to a distinctly higher grade of operation has been effected

out of surplus earnings.

The comparison of gross and net earnings from operation, if

expenditures have grown almost as fast as gross income, confirms the

preceding conclusions. Surveying the chart for the period since 1890,

it appears that net earnings for the railroads of the United States

have more nearly trebled than doubled; the increase having been 177

per cent. up to 1910. This takes no account whatever of the immense

volume of new capital added to the system. The entirely distinct

question of the relative rate of return upon the investment will engage

our attention at a later time. Examination of the years of rapid

revival after 1897 by themselves, however, especially for individual

companies, shows striking results. This is especially true of roads,

not then developed up to a fair working capacity for their plants.

An interesting comparison with the previous decade, 1870 to 1880,

exemplifies this relation still further. The gross earnings of the

trunk lines of the United States decreased very greatly per mile of

line from $7,211 in fact to $6,636 during the decade; but at the same

time the net earnings steadily increased. This was due primarily to the

great volume of business developed,--the ton mileage increasing more

than three fold during these ten years. It happened despite the fact

that the miles of line during the same period had more than doubled.

The following decade, 1880 to 1890, was represented by an increase of

only 82.7 per cent. in mileage, while the number of tons of freight

hauled one mile increased by 132 per cent. Density increasing in this

way, a corresponding ability to carry at a lower rate per ton was a

necessary result. So indisputably has this law--that an expanding

volume of business up to a certain point, may profitably be carried

at a continually lowered cost--been proved, that it is estimated by

so competent an authority as the _Engineering Review_ that, provided

sufficient tonnage be available for 2,000-ton freight train loads, a

cost of one mill per ton mile can be attained. Its significance may be

realized from the fact that the lowest revenue per ton mile reported

for the United States is 2.21 mills per ton mile for the long haul

soft coal business of the Chesapeake & Ohio.[58] This, of course, does

not imply that any railroad in actual operation, carrying all kinds

of freight including a large proportion of local traffic, can in the

immediate future hope to attain this result. It is intended only to

show that, provided the volume of traffic be large enough, the cost of

operation tends to decline as a matter of course, until a condition of

congestion for the existing plant has been reached. At this point a new

cycle of costs of operation and of profits makes its appearance.

The most important single factor in the production of increasing

returns upon a railroad is the density of traffic; that is to say,

the amount of business which can be conducted with a given set of

rails, terminals and rolling stock. In other words, it is the degree

of effective utilization of the plant and equipment. It is too obvious

to need demonstration, after what has been set forth concerning the

nature of railroad expenditures, that economy of operation and,

consequently, profits are more or less directly dependent upon this

fact. Such effective utilization of the property may be secured in two

ways: either by a large tonnage per mile of its line, or else by a

concentration of such traffic as it may have into large train loads,

which can individually be transported at low cost. The first of these

economizes the fixed expenses for roadway and line which respond but

slowly to enlargement of traffic, by distributing them thinly over a

large tonnage; the second economizes the mere movement expenses which

tend to grow less rapidly than the size of the trains. For neither

fuel consumption nor wages of train crews expands _pari passu_ with

the paying load. Fortunate the lot of the railroad which enjoys both

these advantages, of density of traffic per mile of line and of tonnage

capable of such concentration in heavy train units.

Traffic density--the tons of freight carried one mile per mile of

line--is readily computed. The ton mileage, representing the total

transportation service, is merely divided by the number of miles of

line operated. The following graded table illustrates the wide range

of this figure, according to the location of different companies and

the nature of their business, as well as the change in the last few

years.[59]

========================================================================

| 1902-3 | 1910

|--------------------------------------+---------

| |--Percentage of tonnage-- |

| |--------------------------|

---------------------+-----------+--------+--------+--------+---------

Rock Island Company. | 428,116 | 25 | 29 | 14 | 581,000

C. M. & St. P. | 605,139 | 23 | 24 | 17 | 709,000

Great Northern | 657,102 | -- | -- | -- | 814,700

N. Y., N. H. & H. | 802,954 | -- | -- | -- |1,057,000

Wabash | 885,208 | 24 | 32 | 9 |1,322,000

Baltimore & Ohio | 2,181,518 | 6 | 62 | 8 |2,711,000

New York Central | 2,163,000 | 16 | 44 | 12 |2,548,000

Lake Shore ('05) | 3,355,209 | -- | -- | -- |3,911,000

Penn'a Railroad ('05)| 6,337,625 | -- | -- | -- |5,139,000

======================================================================

The first of these companies operates in a sparsely settled

agricultural territory. The St. Paul system lies nearer Chicago, but is

still largely dependent upon a local and rural constituency. The Great

Northern--a great transcontinental trunk line--despite its sparsely

settled western area, exchanges a large volume of through freight for

the Pacific Coast for lumber and other bulky products carried east.

The New Haven serves perhaps the most densely settled area in the

United States, but much of its traffic is on branch lines and is of a

retail character. The Wabash lies in well settled territory and hauls

a heavy tonnage of low-grade freight. The last two are not only great

trunk lines to the seaboard, but also tap the coal and iron fields.

Much of their tonnage is consequently of low grade. The Pennsylvania

enjoys a still further advantage, super-adding a rich local traffic in

manufactures and merchandise. As compared with its rival trunk line,

the New York Central, it hauls relatively little grain; but, on the

other hand, the New York Central has a much smaller coal and iron

business. Some one has aptly characterized the difference between

the two roads, describing the New York Central as "operating between

good points, but not through a good country" so far as local business

is concerned. On the one, through traffic is supplementary to local

business, while on the other it is the reverse. The high density of

trunk line traffic is such that about two-thirds of all the tonnage of

the United States is transported east of the Mississippi and north of

the Ohio river.

Traffic density has enormously increased during the last two decades,

as a result of the filling up of the country and the relative cessation

of new construction. This is manifested by the growth since 1890. In

that year the density was less than 500,000 ton miles per mile of line,

and during the depression of 1893 it fell well below that figure. The

total of 1,053,000 reported for 1911 represents, therefore, more than

a doubling of the density in twenty years. This growth during 1898 and

1905-'06 was notable. The latter period, especially, was a time when

congestion upon all the roads of the country occasioned much distress.

The fact is evident that the country has well grown up to the measure

of its existing transportation facilities.

The second measure of effective operation for the production of

increasing returns, is concentration in the trainload. This is regarded

by many as the supreme test of efficiency in management. Great progress

has been made during the past years in this regard in the United

States--an improvement which has largely enabled the carriers to

bear up under an increasing burden of expenditure. The trainload is

generally adopted today as the unit of operation, measuring the cost

of service.[60] It is a fact that, within certain limits, the cost of

handling a train does not vary greatly with its capacity. Since the

first application of air-brakes to freight trains in 1887, a train

crew sufficient to handle fifteen cars can care for thirty about as

well in long haul wholesale business. Fuel cost, also, as has been

shown, lags well behind the rate of increase of the load. Eaton in his

Railroad Operations, concludes that from thirty to fifty per cent. of

cost is independent of the trainload. The effect is that any increment

in the paying load very materially decreases the cost of operation per

ton.

Progress in the United States in increasing the average train load

is shown by the lowest curve on the diagram on page 97. The scale

applicable is along the left hand side of the chart. From 175 tons

per train in 1890 to an average figure of 383.10 in 1911 is certainly

a remarkable showing. The most rapid increase seems to have occurred

after 1897, with the first resumption of general prosperity. As for

individual roads, the following table of average train loads is

suggestive, as showing the gradation between roads of different type,

as well as progress from year to year:

AVERAGE NUMBER OF TONS OF FREIGHT PER TRAIN (TONS PER TRAIN MILE)

==============================================================

Road (Fiscal Years) | 1901 | 1905 | 1910

-----------------------------------------+------+------+------

Pennsylvania Railroad (East of Pittsburg)| 478 | 498 | 649

Pennsylvania Company (West of Pittsburg) | 382 | 420 | 511

Pennsylvania System (Both) | 454 | 476 | 607

Chesapeake & Ohio | 511 | 557 | 701

Great Northern | 347 | 541 | 520

Erie Railroad | 379 | 416 | 497

New York Central & Hudson River | 365 | 381 | 413

Northern Pacific | 324 | 367 | 429

Atchison, Topeka & Santa Fe | 238 | 265 | 298

Chicago & Northwestern | 255 | 259 | 261

New Haven | 208 | 222 | 293

Southern | 190 | 194 | 237

==============================================================

The great coal and iron roads, the trunk lines and the transcontinental

lines all concentrate their business; while the granger roads, like

the Atchison and North Western, the roads with much local business

like the New Haven or the Southern, operating in sparsely settled

regions, all have of necessity smaller trainloads. But all alike betray

remarkable progress in this regard. In 1870 the average for the best

roads was little above 100 tons,--such as 103 tons reported for the New

York Central and 137 tons on the Lake Shore. From this level to results

of 400 or 500 tons on the average represents a notable achievement. The

Lake Shore for 1911 reports a revenue train load of 635 tons. It should

be observed, however, that such results come from longer trains, not,

apparently, so much from larger cars. To raise the average trainload on

the Wabash from 196 tons in 1890 to 386 tons in 1908 is also worthy of

note. The significance of these recent figures can be realized from the

fact that the London & North Western, one of the leading railroads in

Great Britain, reports recently an average freight train load of only

68 tons. This represents probably a fair average for European railroads

as a whole, although in England the general practice of privately owned

cars, of light locomotives, short freight sidings, etc., may reduce

the figure slightly below the continental average. Statistics not only

show the notable improvement in recent years; they at the same time

show how the trainload performance is affected by trade conditions. For

nearly every road the trainloads for 1904 were distinctly lower than

in the preceding years. This was a year of acute business depression.

The movement of great staple commodities, such as iron ore, coal, steel

and iron and lumber, was greatly curtailed. All business was conducted

on a narrower basis. Smaller trainloads were an almost inevitable

consequence. The revival in the following year, however, immediately

improved the conditions of operation, as the figures indicate.

It will be noted that the figures for the American roads above given

represent averages. These are compounded from local and through traffic

taken together. It is apparent at once that local trains must average

far lighter loads than are customary upon long hauls without breaking

bulk. Thus New England railroads report for 1906 an average trainload

of only 220 tons, while other parts of the country, such as the North

Central group, report 426 tons of paying load. Only by separation of

local from through business can we adequately appreciate the enormous

advances which have taken place in railroad operation in the United

States, with corresponding reductions in the cost of transportation.

While the New York Central at one time reported an average trainload

of 322 tons, the average load of its through trains on the main line

rose as high as 750 tons. More than twice this figure is attained upon

the Pittsburg, Bessemer & Lake Erie road in hauling ore from the lakes

to the furnaces at Pittsburg. The Illinois Central, for its low grade

and long haul to the Gulf, has recently built locomotives capable of

hauling 2,000 tons of net paying load. A standard grain train on the

Lake Shore in 1903 consisted of fifty cars holding forty tons each.

Even this figure has recently been surpassed by the New York Central,

which, with its monster new "mogul" engines, hauls eighty loaded 30-ton

cars, giving 2,400 tons of revenue freight. The Mallet locomotives with

a tractive effort of 100,000 lbs., at present seem to have reached the

limit of size and weight. Seventy-five grain cars holding 1,000 bushels

apiece are equivalent to the production of twenty bushels per acre of

an area of six square miles. This is an ordinary trainload. It is not

infrequent to transport a fifth more than this. Eighty and even one

hundred cars in a train since 1900 often bring the load up to 3,600

and even 4,000 tons of freight. Such a train is over four-fifths of a

mile long. From these figures it certainly appears that trainloads for

long haul are standardized at not less than 2,000 tons, a figure which

would have seemed absolutely impossible to railroad managers of fifteen

or twenty years ago. The maximum trainload in Germany on coal traffic,

which, of course, greatly exceeds any general average for trains of all

classes, is about five hundred tons. It has been regarded as a notable

achievement that this represents an increase of about one hundred tons

in the last decade.

On the other hand, the extravagant promises of economy from large

trainloads have been considerably abated of late. It has been

effectively demonstrated that there is a limit to such growth. Only

low-grade and long haul carload traffic can profitably be concentrated.

In 1903, for instance, a general decrease in trainloads followed a

reduction in the relative amounts of low as compared with high grade

tonnage. Less iron, coal, and raw materials and more merchandise and

manufactures offered for carriage, necessitated a positive reduction

in the trainloads as already mentioned. Nor can local business in

less than carload lots profitably be concentrated beyond a certain

point. Grades must be uniform to attain such economy. The trainload

must not exceed the traction power on the heaviest inclines, or else

expensive pusher engines or breaking up of trains will offset all other

savings. Moreover, too great trainloads even on the best roadbeds

involve slower speeds. Not only is other traffic thus impeded, but

the economy in wages vanishes after a certain point with such slower

movement. The fashion had been set by James J. Hill, the master mind

in the transcontinental field. His notable results, due to a careful

working out of every detail, led to a frenzied imitation on all sides.

Many roads then discovered to their loss that while they had provided

rolling stock for heavy loading, ampler terminals, longer sidings

and heavier bridges also were a necessary accompaniment. Part of the

congestion of traffic in 1906-1907, already mentioned, and a portion

of the financial embarrassments of recent years, were undoubtedly due

to too great haste in seeking economies of this sort in rolling stock,

without at the same time making provisions for enlargement of other

portions of the plant. A more discriminating policy has consequently

resulted of late. Traffic is being sorted according to its availability

for concentration. The best utilization of the rails and terminals is

being more considered. Business demands for quick delivery also enter

into the calculation. Instead of a few huge slow-moving leviathans

blocking other trains, the line may perhaps better be kept full of many

smaller trains moving more nearly together. Such are certain of the

details now being worked out. None of them, however, weaken the main

proposition that a discriminating concentration of traffic conduces

very greatly to economy of operation.

This concentration of traffic units is largely due to technical

improvements of various kinds. Foremost among these has been the

development of the steel rail. In 1880 more than seven-tenths of our

mileage was still equipped with iron rails. Rapid progress ensued

during the next ten years, upward of eighty per cent. being in

steel rails by 1890. At the present time, the proportion is above

ninety-eight per cent. In fact, no iron rails have been made for many

years, except for repairs and on insignificant branch lines in remote

parts of the country. A steady increase in the weight of the rails has

ensued. The standard rail for main lines until the Civil War weighed

fifty-six pounds to the yard. In the seventies this was increased to

sixty-three and above; in the latter eighties the best practice was to

use seventy-five pound sections. Since 1900, they frequently run as

high as one hundred pounds, in regions of dense traffic. Few main lines

of track now average less than seventy-five pounds. It is this increase

in the use and size of steel rails which has permitted improvements in

rolling stock. But, on the other hand, grave dissatisfaction with the

quality of the rails manufactured of late years, particularly since

the establishment of practical monopoly under the United States Steel

Corporation has become manifest. Numerous accidents due to breakage

of rails, especially since 1905, have revealed either defects in

manufacture or an undue load imposed by heavier rolling stock, too

high speed, or both. The matter has become steadily worse. In 1902 the

Interstate Commerce Commission ascribed seventy-eight accidents to

broken rails. Nine years later the number had risen to 249. The need

of improvement is now fully recognized on all sides.

The power and efficiency of locomotives has increased, perhaps, more

since 1890, and particularly since 1895, than in any previous period.

Superior materials particularly have contributed to this result, such

as the substitution of cast steel for cast iron and of nickel steel for

wrought iron in axles, crank pins, etc. Some of the improvements which

may be mentioned are, for instance, an increase in the average heating

surface from 2,000 in 1890 to nearly 3,000 square feet at the present

time, and an increase in the average steam pressure from 160 pounds

to 210 pounds per square inch in the same period. The maximum weight

has also increased very rapidly. The average weight of a locomotive

at the close of the Civil War was approximately 90,000 pounds. This

has increased in somewhat the following proportions: To 1881, 102,000

lbs.; to 1893, 135,000 lbs.; to 1895, 148,000 lbs.; to 1898, 230,000

lbs.; rising in 1900 to 250,000 lbs. Passenger locomotives since 1892

have almost doubled in weight, and freight engines have more than done

so. Compound and double or Mallet locomotives are also supplanting

those of simpler type for peculiarly heavy service. The first compound

engine was built in 1899, only one being constructed in that year. In

1900 a single locomotive works turned out 500--a number constituting

two-thirds of the entire output of that company--for use in the United

States. Such locomotives cost more in first instance; but the greater

weight and steam capacity, together with the considerable saving in

fuel, amounting to perhaps twenty per cent., more than offset this

objection. The traction efficiency of these improved locomotives may be

shown by the statement that in 1885 the decapod Baldwin locomotives,

made to haul 3,600 tons on a level, represented the maximum capacity.

Five years later the same company built locomotives to haul 4,000 tons,

not only on a level, but on any ordinary grade. As indicative of late

advances in locomotive construction, we may instance those built about

1900 for the Illinois Central and the Union Railroad at Pittsburg,

both low-grade roads, carrying exceedingly heavy train loads. The

first of these weighed, including its tender, 365,000 lbs., the Union

Railroad consolidation engines weighing 334,000 lbs. Such locomotives

are stated to be twice as powerful as the best which were manufactured

twenty-five years ago. This record is surpassed by engines which have

just been built for pusher service on the soft-coal Virginian Railway.

They are of the Mallet type, weighing 540,000 pounds; with a train

capacity of 4,230 tons. The evaporative surface is 6,760 square feet.

As summarizing the increased efficiency of American locomotives, we may

instance the figures of the Interstate Commerce Commission, showing

the average performance of locomotives for the United States. Whereas

in 1894 the average number of tons of freight carried per locomotive

was about 32,000 tons, this rose to 46,000 tons in 1899, and to 54,600

tons in 1906. At the same time the number of tons of freight hauled one

mile for each freight locomotive in the United States increased from

4,000,000 in 1894 to approximately 6,000,000 in 1899, and to 7,300,000

in 1910. In other words, the average performance of each freight

locomotive in the United States has increased by more than fifty per

cent. in the last decade.

The economy of large freight cars has been amply demonstrated. Marked

advances in the average capacity have taken place in the last few

years. In the sixties a 15,000 pound freight car represented about

the normal capacity. This has increased, as measured by maximum load,

to 28,000 pounds in 1873; to 40,000 pounds in 1875; to 60,000 pounds

in 1885; to 70,000 pounds in 1895; while at the present time 80,000

to 100,000 pound cars are in everyday use. Cars of this latter type,

built to carry forty to fifty tons, are necessarily of pressed steel

construction, and are mainly useful for the carriage of coal and

ore and similar low-grade commodities. It seems to be questionable

whether a maximum capacity has not been about reached, in view of

the exceedingly great wear and tear imposed upon track, bridges, etc.

Up to this point the economy of heavy loading is indisputably proved.

Increased size of cars far more than proportionately increases the

paying load. Thus, for instance, an 18,000-pound car will carry 20,000

pounds load, while a 22,000-pound car will carry a load twice as great.

It is stated on good authority, for example, that a car of forty tons

capacity can be built which will weigh but 3,000 pounds more than a

thirty-ton car, and cost hardly fifty dollars more. This is undoubtedly

the reason why at the present time the average load per car is at least

one hundred per cent. greater than the maximum which was possible

twenty years ago.

A steady increase in the freight performance of American equipment

is shown by official data of the Interstate Commerce Commission.

Whereas in 1894 it required on an average 1,888 freight cars for every

1,000,000 tons of freight transported their capacity has so increased

that the same amount of traffic in 1906 was carried by only 1,127 cars.

In other words, an enormous increase in the freight service had been

attained. On the other hand, the actual mileage performance of much of

this equipment is extraordinarily low. It averages only about 9,000

miles annually or an equivalent of thirty miles a day. At a speed of

fifteen miles an hour, this means that actual movement under a paying

load, allowing for one-third of its journeyings empty, occupies but

little over an hour and a quarter a day. The actual performance is,

however, not quite as poor as appears. For, of course, this average

includes the non-movement of all cars in bad order (sometimes one-tenth

of the total); and also all idle equipment. This latter consideration

is of great moment. Special cars, suitable only for seasonal business;

and especially demurrage delays, often forty-eight hours or more,

adversely affect the result. Where separate mileage records of

"foreign" cars are kept, as on the Wabash system, it appears that

their mileage is twice as high as for "home" cars. The difference is

due to the fact that cars off their own rails, mainly are in actual

demand and are kept moving. Probably the daily performance of loaded

cars is not less than 150 miles. But a journey of this length, with two

days' delay at each end at terminals, would bring the average down to

about thirty miles. The public does not always appreciate these facts;

and is often querulous. It is certain that the problem how to secure

greater efficiency in the use of this equipment is as yet imperfectly

solved.[61]

[Illustration: Revenue and Cost Per Train Mile.]

The discussion of the nature of railroad expenditure may be concluded

by a comparison of the net effects of the developments of the last few

years; that is to say, of steadily expanding costs of operation and of

slowly and tardily rising rates chargeable for service on the one hand,

as over against the results obtained by mechanical improvements and

increasing economy of operation coupled with growth of tonnage, on the

other. The average cost of transportation has greatly increased. This,

according to the statistics of the Interstate Commerce Commission, is

shown upon the diagram herewith by the middle curve.[62] The average

cost of running all trains per mile, which had fallen from 96 cents

in 1891 to 91.8 cents in 1895, rose to $1.07 in 1900, and in 1911

increased by more than one-third, to $1.54 per mile. Against this

should be set the fact that while the trains thus cost more to haul per

mile, their paying load has increased in somewhat smaller proportion.

This is shown by the upper curve on the diagram above mentioned.

For freight trains the increase has been from $1.65 to $2.89 per

mile. Passenger revenues per train mile have increased less rapidly.

This follows from the well-known fact that freight rates have been

increased, while passenger rates have not changed for the better during

this period; and also that economies in concentration of traffic are

necessarily confined to the carriage of freight. The immense gain in

trainloads has probably been the main element, among these, as already

observed.

Among other things this diagram also brings out the effect upon

revenue of the substantial rate increases after 1900, coupled with the

elimination of rebate losses. It will be observed how sharply the upper

curve of revenue per train mile slants upward after 1899, by comparison

with the lower line denoting cost. The same thing apparently occurs

again after the set-back of 1907.

The interrelation between these various factors may be more readily

shown by confining our attention to the period during which a

practically uninterrupted development of business ensued, thus

eliminating the confusion due to the four years of depression

after 1893. The data on our various charts for the years 1898-1906

demonstrate that during this period the ton mileage, measuring the

freight traffic handled, has practically doubled. To transport this

doubled tonnage, a growth in freight train mileage of only eighteen

per cent. was necessary. This was due, of course, to the notable

concentration of train loading, already described, as well as to a

density of traffic per mile of line almost sixty per cent. greater. As

a consequence of these economies in operation, the revenue per freight

train mile has increased by about fifty per cent.; while the average

cost of running all trains per mile has grown less rapidly, namely, by

42 per cent. Had we data for freight trains alone it would surely be

lower than this. In the meantime during this period of eight years, the

rate of return in revenue per ton mile received, remained practically

unchanged.[63] From all of which it would appear that even despite all

these confusing factors, the law of increasing returns, so far at least

as 1898-'06 was concerned, was making itself appreciably felt.

Attentive consideration of the available figures, especially as shown

by diagram on page 97, shows apparently that the various economies in

operation, heavier trainloads and the like, have not since 1906 yielded

any greater profit from mere operation, with the ever increasing volume

of business. In other words, the increase in the margin between cost

of operation and revenue per train mile,--measuring profitableness

per unit of movement--has not kept pace with the augmentation of the

size of that unit,--the trainload. Thus it follows, as one would

expect, even making allowance for all changes in rates, wages and

other expenses, that the law of increasing returns as applied to

railroads, does not arise primarily from economic considerations

as to mere physical operation. The law originates primarily in the

fiscal conditions attaching to the heavy capital investment,--the

fact, namely, that fixed charges up to a given point of saturation

tend to remain constant, absolutely; but become proportionately

less, therefore, as the volume of business expands. From this fact,

therefore, rather than because of any marked economies of large-scale

production, may it be affirmed that railroads offer a notable example

of the law of increasing returns. The important bearing of this

distinction will appear in due time in connection with the problem of

the determination of reasonable rates. Added significance, also, is

given to the relation between the cost of new capital, measured by the

rates of interest on bonds and dividends on stocks, and the supply

necessary to provide adequate extensions and improvements in future.

_Appendix_:

The subjoined chart, reproduced by the _Railway Age Gazette_ from a

bulletin of the Bureau of Railway Economics, brings out forcibly the

manner in which, within the short time limits of full utilization

of plant, a large increase of business can take place without a

commensurate growth of expenses. The phenomenon for railroads is of

course cyclical. Annually, as here indicated for 1911, the second

half of the year is marked by a much heavier movement of traffic,

principally, of course, the crops. But expenses never rise in

proportion. This is most evident for the eastern group of roads, as

here shown. This causes the net revenue curve, also, to vary much more

than in proportion to the volume of traffic, in consequence.

[Illustration: Monthly Revenues and Expenses per Mile of Line from 1911

and 1912.]

FOOTNOTES:

[50] Illustrated by the seasonal variation of business. _Vide_ appendix

to this chapter at p. 100.

[51] Webb, Economics of Railway Construction; originally in

Wellington's Economic Theory of the Location of Railways.

[52] Webb, _op. cit._; originally from Wellington.

[53] _Wall Street Journal_, March 25, 1908.

[54] Eaton, Railroad Operations, etc., pp. 44-58.

[55] Investigation in 12th Ann. I.C.C. Rep., 561.

[56] From Report of Commission to Investigate the Postal Service, 1901,

p. 220; brought down to 1906 when local disturbances in wages, other

costs of operation and rates outweigh all general considerations.

[57] Between 1890 and 1910 freight ton mileage rose three times over.

Operating expenses grew by about two and one-quarter times.

[58] _Cf._ data as to revenue per ton mile on p. 413, _infra_.

[59] Other data as to density on p. 413, _infra_.

[60] Also known as "average tons per train mile." Obtained by dividing

the ton mileage by the sum of the freight and mixed train miles.

[61] _Cf. Quarterly Journal of Economics_, XVIII, p. 299; on _per diem_

reform. Also, _Railway Age_, 1903, p. 136; 15th Ann. Rep., I. C. C., p.

79; and Circular Letters, 1901, Chicago Bureau of Car Performances.

[62] Using the right hand scale.

[63] Diagram on p. 413, _infra_.

CHAPTER IV

RATE MAKING IN PRACTICE

Evolution of rate sheets, 101.--Terminal v. haulage

costs, 102.--Local competition, 104.--What the traffic

will bear, 107.--Trunk line rate system, 111.--Complexity

of rate structure, 113.--Competition of

routes, 114.--Competition of facilities, 116.--Competition

of markets, 118.--Ever-widening markets, 119.--Primary and

secondary market competition, 121.--Jobbing or distributive

business, 124.--Flat rates, 127.--Mississippi-Missouri

rate scheme, 128.--Relation between raw materials

and finished products, 134.--Export rates on

wheat and flour, 135.--Cattle and packing-house

products, 139.-- Refrigerator cars, 140.--By-products

and substitution, 142.--Kansas corn and Minnesota

flour, 143.--Ex-Lake grain rates, 145.--

The task of constructing a freight or passenger tariff is an eminently

practical one. The process must be tentative and experimental. Little

can be calculated in advance. Tariffs are not made out of hand; they

grow. Not until a rate has been put into effect, can its results be

known. The lower limit of charges, however, is more or less fixed.

Obviously the rate must not be less than that portion of the variable

expenses incident to each particular unit of business. This variable

expense is divisible into two parts, one for loading and unloading,

and the other for actual movement. The first step in constructing a

tariff, therefore, is to separate these two portions of the variable

outgo. General experience fixes the terminal outlay for loading and

unloading at an average figure of about twenty or twenty-five cents per

ton at each end of the line; that is to say, at an average of about two

and one-half cents per hundred pounds as the total terminal cost.[64]

Just where, above or below this average, the figure for any particular

tariff will lie, depends upon a multitude of details.

This terminal expense is obviously quite independent of the length of

the haul. It costs no more to load for a carriage of 3,000 miles than

for one between two adjoining stations. It is the second portion of

the specific costs, namely, the movement expense, which varies with

the distance. This movement cost is more difficult of determination,

as affected by a multitude of variable factors, such as the grades,

curvature, number of stops, the size of train load, and above all,

the volume of the traffic. Assuming the simplest physical conditions,

one would expect the movement expense, aside from the initial cost

of getting up steam in order to move at all, to rise proportionately

to the distance traversed. Graphically represented, the tariff would

appear somewhat as follows:

[Illustration: Relation of Cost of Carriage to Distance.]

In this diagram the distances of carriage are represented along the

horizontal line, A B; while the rate charged is laid off vertically.

The distances A C and E B represent the constant terminal cost; while

the steadily rising rates with increasing distance, due to movement

expenses, are shown by the sloping dotted line C D. This chart at

once demonstrates why under the very simplest physical conditions a

straight mileage tariff is unscientific and unreasonable. For the

constant terminal expense, spread evenly over the mileage traversed

as the movement expenses grow, becomes progressively less and less

in proportion to the total of the two, which constitutes the real

rate. The longer the haul, the lower the ton-mile cost as a matter of

necessity. As Chanute calculated on the New York Central a generation

ago,[65] while the average cost per mile of hauling a ton ten miles

was 4.062 cents, it descended progressively to less than one cent per

mile for distances over five hundred miles. A common rule is that

the rate rises as the square root of the distance, rather than in

proportion to it. A hundred-mile haul represents a cost approximately

only twice as great as one of twenty-five miles, instead of being four

times as much. For thrice a given cost the haul may be increased nine

times. The course of such a tariff with increasing distance would be

represented by the parabolic curved lines on the preceding diagram.[66]

The particular curve would depend upon the commodity and local physical

conditions. In territory where movement expenses were heavy or

operation difficult, the curve would obviously rise more rapidly. Such

a mathematical tariff does not depart widely from the one traced by

the heavy dotted line C D first described. The progressive decline of

the per mile rate with increasing distance may be illustrated by the

rough estimate of allowing two and one-half cents per hundred weight or

fifty cents per ton for terminal cost, with one-half cent additional

per mile for movement expenses. For a ten-mile haul this would cost

fifty-five cents, or an average of 5.5 cents per mile. Were the

distance 500 miles, the average cost would be only (50+250)/500 cents

or 0.6 cents per ton mile.

[Illustration: Diagram of Belgian Tariff Sheets.]

Thus far the problem has been seemingly simple. The next step

introduces new complications. Our hypothetical railway line at a point

one hundred miles out, may cross a navigable river or canal, or may

intersect another railway. Engineering considerations of absolute cost

of operation now no longer predominate. Relative costs by rival lines

enter into the case. Water lines or more direct railways compete for

the traffic. One cannot even fall back upon the cost of carriage by

any of these lines, either the weaker or the stronger. An entirely

new principle comes into play. The alternative is presented of taking

the business at a rate lower than, and out of line with, rates on

general traffic, rather than to lose it to another line. At first

sight it would appear that it were better to abandon the traffic than

to take it for less than a fair average return or profit. This is a

serious matter. The tonnage offered is large. The existence of active

competition for it, is proof of its importance. Railways meet at large

towns, and large towns become larger because the roads meet there. The

main reason for not abandoning the traffic, however, arises from that

primary fact, to which one constantly recurs, that all expenses are not

alike in their nature. A concrete example will make this plain.

Suppose, for instance, the normal rate to yield a fair average return,

all expenses considered, be thirty cents per hundredweight. Two-thirds

of the cost of this, or twenty cents, would not cease as outgo, were

this business abandoned. The rails would rust, the ties would rot, and

trains would move but with lighter loads, and the fixed charges would

still go on inexorably night and day. Ten cents per hundredweight will

meet the variable and extra cost incident to this particular business.

A fifteen-cent rate would at least repay these extra outlays. It

would do more. It would contribute five cents per hundred pounds to

the twenty cents outgo per hundredweight, which, without the traffic,

would have to be borne _in toto_. Even a rate of eleven cents would

contribute something to this end. For it would leave a surplus of one

cent per hundredweight to lighten the other burden. Adopting Hadley's

phraseology,[67] if you take at eleven cents, freight that costs you

thirty cents to handle, you lose nineteen cents on every hundredweight.

If you refuse to take it at that rate, you lose twenty cents on every

hundredweight you do _not_ carry. For your constant expenses go on,

while the other road gets the business. There is only one course open.

The rate at the competitive point must be cut; if not to make a profit,

at least to stop a greater loss. And one comfort may be uncovered in

so doing. The lowered rate may so stimulate new business and enlarge

the volume of traffic, that it may be handled at much lower cost. In

fact, this consideration alone in absence of all competition, may

induce a lowering of rates at certain points out of line with the

general schedule. This incentive, conditioned by the fact of increasing

returns, is always in the background. The destiny of many places is

manifested in terms beyond the control of the carrier. Soil may be

poor, climate or population adverse to progress. But some particular

places enjoy peculiar advantages for growth. Not to stimulate new

business at these points where traffic might be cultivated, even

without rivals in the field, is little better than allowing it to

escape over a competitor's line of rails, were they present.

[Illustration: Effect of Competition at Certain Places on Rates.]

Cutting the normal rate at competitive points or at important points

in order to stimulate traffic, in conformity with the principle above

stated, transforms our tariff diagram as shown herewith. The rate

rises steadily with the increasing distance from A, except at E and F.

At these points it is fixed at a lower point, determined not primarily

by the cost of service at all, but by the available demand for it.

Traffic at these points is charged what it will bear; not as much but

as little as it will bear: which, being translated, means that the

charge is set as high as possible, still holding the volume of business

constant, or even increasing it if that can be accomplished. The total

profit is constituted of the profit per unit of freight multiplied into

its volume. The centre of interest is here shifted from the average

profit per unit considered alone, to the total profit thus obtained.

At this point another difficulty presents itself. Although, as set

forth elsewhere, local discrimination,--charging a lower rate for

a more distant point,--may sometimes not only be not injurious but

actually beneficial to all parties concerned, it is the exception, not

the rule.[68] Ordinarily to accord a remote point a lower rate without

patent cause, is an economic anomaly, and, moreover, a political

blunder. It violates the democratic principle of cost of service as

underlying rate schedules. Most legislative bodies have prohibited it

by law. The United States and most of the American commonwealths do not

permit it, other than in exceptional cases. The result is that on our

hypothetical tariff, the rates from A to points intermediate between A

and B and B and D must be cut to the levels, E and F, fixed for those

latter places. Such was the action taken by the trunk lines in 1887 in

conformity with the requirements of the long and short haul clause of

the Federal Act to Regulate Commerce. An original progressively rising

tariff is thus at once transformed to a series of level grades or

platforms, the shifts of level corresponding to the location of large

towns or competitive centres; and the grade of each platform being

fixed by the rate determined under competition at those points.[69]

This ascending series of grades may be most irregular, as conditioned

by local circumstances. The general steepness of the gradation is low

on eastern roads like the New York Central, with a large volume of

traffic and easy operating conditions. On western lines like the Denver

and Rio Grande, in rugged territory, with a sparse population and light

tonnage, the per mile rate rises rapidly and the gradation of the

general tariff is steep. But always it will be found that the changes

in rates occur at competitive points, with transition to a new level of

rates determined by the conditions at the next competitive point beyond.

An important fact concerning this tariff thus far developed, is that,

of course, the height of the upper level at the most remote point must

never exceed what the particular traffic will bear. In other words,

supposing that the traffic consist of grain or coal, not more than a

certain amount could ever be charged, no matter how great the distance,

without so far diminishing the profit in the transaction as to render

the business impossible. This is shown by the diagram opposite the next

page, whereon it appears that each commodity, coal, wheat, cement,

lumber, or oil, having attained a certain level of rates, never rises

thereafter, no matter what the distance. Each attains the maximum of

what it will bear. That level it can never exceed. This immediately

leads to another consideration. No single tariff is applicable to

any large number of commodities. Each one must be regarded as a law

unto itself. Not only does the ultimate amount which each is able to

bear depend upon the value of that commodity, but also the conditions

determining competition with respect to it must be different all along

the line.[70]

[Illustration: RATES BETWEEN CHICAGO AND ST. PAUL. DISTANCES IN MILES

FROM CHICAGO.

RATES IN CENTS PER 100 LBS.

[_Facing page 108_]]

Thus it appears that the height of the extreme upper level in our

diagrammatic series of rates is fixed by the highest charge which that

particular traffic will bear.[71] Beyond a certain point, no matter

how great the distance, the rate cannot be increased above this level.

This maximum varies, of course, with each commodity. On cotton it may

be fifty-five cents per one hundred pounds; on grain or coal it will

be much lower, and on sand or cement lower still. The problem of the

traffic manager is to attain this highest rate as speedily as possible

with increasing distance, and to grade his rates with distance up

to this level as quickly as possible, consistent, of course, with

maintenance of a full volume of business. But not only may the final

limit of what the traffic will bear be different for each commodity;

the steps or stages by which the rate progresses up to this maximum,

are quite independently determined. The actual tariffs of local class

rates in general are much simpler than the commercial conditions of

rate making often warrant. Probably the major portion of tonnage on

American railways moves under special or commodity rates. Even in

Prussia over three-fifths of the traffic is of this exceptional sort.

These special rates are made with a view to particular circumstances

prevalent at the time. Bids from a quarryman in Vermont on stone for

a public building in Chicago, may be dependent upon the grant of a

low rate on his marble in competition with a quarry in North Carolina,

also able to supply the particular stone required. The various

ascending series of rates are thus rendered bewilderingly complex.

This is also shown by the foregoing diagram of rates between St. Paul

and Chicago.[72] The rate on a cheap, heavy commodity like coal,

probably rises rapidly at first, and soon attains a maximum beyond

which it can never go. On this diagram, for instance, the freight rate

on soft coal for points up to 180 miles out is lower than that on

flour. Beyond that point the coal rate in turn exceeds that on flour.

Cement is higher than lumber for the first 150 miles; but after that

point the relatively greater value of lumber holds it steadily above

cement. On heavy cheap commodities the relatively high cost of cartage

in competition enables the railway to reap the full measure of its

advantage and to charge well up to the maximum of what the traffic

will bear, within a comfortably short distance. Furthermore, variable

costs for terminal charges have to be considered. Wherever they are

high the rate must rise at once sufficiently to cover these, no matter

how short the distance; but thereafter the rate may not need to be

increased greatly for some time. On light higher-grade goods the wagon

is an effective competitor for longer distances.[73] Moreover, the

competitive points at which rates rise from stage to stage are seldom

the same for all classes of goods. A river crossing brings competition

for coal, lime, or cement, but does not affect the rates chargeable on

high-class freight which seldom goes by water in any event. A railway

specially interested in the development of some particular industry,

wherever it crosses our hypothetical line, effectively holds down

the rate on the product of that business. Junction points with other

railways having no such interest may have no influence upon that rate,

but may cause modifications in other directions. Another railway being

in need of back loads over its line, as the result of a predominant

movement, let us say, of beef cattle at certain seasons of the year,

may introduce competition in all the tonnage capable of being carried

on cattle cars. Such a road holds down the rates on this traffic

wherever it happens to cross, but has no effect upon any other rate.

Thus it comes about in practice, as the last diagram well illustrates,

that the tariff lines cross and recross one another, generally rising

with increasing distance, but at all sorts of different times and

places.

Few generalizations are possible in this connection. Rate making

must in a growing country ever be a matter of infinite detail. It is

generally true, however, that beyond a certain point the tariff on

different grades of commodities will separate more and more widely

with increasing distance. For, obviously, after the low-grade goods

have reached the maximum which they can bear--and this they tend to do

speedily--they must remain practically constant; while those of higher

grade continue progressively rising. And for very short distances

the rate on the low-grade goods may even exceed that imposed upon

higher-class tonnage. The coal rate for a ten-mile haul may exceed that

upon some commodity worth twice as much; but for a 200-mile haul the

coal rate may be only one-eighth of the rate on the other goods. Long

experience on the part of the carriers has, however, enabled them to

arrange their tonnage in classes; for each of which the conditions are

more or less uniform. By reserving the exceptional traffic for special

treatment under commodity rates, a fairly consistent scheme of charges,

rising by stages with increasing distance may be evolved.

Few standard railway tariffs in the United States develop beyond the

point covered by the preceding paragraph. Many of them are unable even

to reach this stage of logical growth. In the South, for instance, they

have never got beyond the stage of progressively rising local rates,

with independent and often radically reduced charges at all large towns

or competitive points.[74] Each traffic manager, particularly since

the effective prohibition of working agreements between competing lines

by the Trans-Missouri Freight Association decision of the Supreme Court

in 1896, has been left to work out his own salvation, not aided by, but

in spite of, the efforts of his rivals. There is, nevertheless, one

example of further development in the so-called trunk line territory,

lying east of the Mississippi and north of the Ohio and Potomac rivers.

Conditions here, in general, are most favorable by comparison with the

West and South. Both population and traffic are dense, and the state

legislatures are conservative in making grants for the construction

of new lines. The companies are historically mature. The good fruits

of coöperation had already appeared in the evolution of a scientific

and logical scheme, long before such coöperative action had been

frowned upon by the law and the courts. All the railways in trunk line

territory have worked in harmony, so far as general classified local

tariffs are concerned--however much they may have fought one another

over differentials to seaboard cities, or export and import rates.

Their system is comparatively simple in principle, although it has

required the experience of many years to work out in detail. Fully

described elsewhere,[75] it will suffice for present purposes to say

that all rates from intermediate points between Chicago and New York,

are fixed at a definite proportion of the Chicago-New York rate both

for east-and westbound shipments. Thus, for instance, as shown by the

map of trunk line rate distribution, at page 365, the rate from Detroit

to New York is seventy-two per cent. of the Chicago-New York rate.

The percentages from the following points are as indicated, namely:

Cincinnati, eighty-seven per cent.; Indianapolis, ninety-three per

cent.; Grand Rapids, ninety-six per cent.; Peoria, Ill., one hundred

and ten per cent.; Louisville, Ky., one hundred per cent.; Milwaukee,

one hundred per cent.; and even points in Canada, such as Toronto,

seventy-eight per cent., etc. Every place, no matter how small, has

a certain percentage of the New York-Chicago rate assigned to it.

This rate changes with any variation of the standard or basic charge.

Thus when the Chicago-New York rate, first-class, is seventy-five

cents, the rate from Indianapolis is ninety-three per cent. of that

figure. Any change of Chicago-New York first-class rates modifies

every intermediate rate in exactly the same proportion. This was well

exemplified in the rate wars of 1893. These percentages have been

fixed after a long process of compromise among conflicting interests.

Another point of special interest is that these rates are adjusted on

the basis strictly of the long and short haul principle, namely, that

all intermediate points enjoy a somewhat lower rate than the terminal

points, although the percentage may not be exactly upon a mileage

basis. Consideration of the distribution of these percentages points

to many apparent inequalities in the adjustment; but, as a matter of

fact, it will be found that the existence of competing routes, of water

transportation or of other factors, offers a partial explanation in

most instances.

[Illustration]

Such being the general character of this comprehensive trunk line

system, the relation of it to the tariffs described heretofore is not

difficult to demonstrate. Each separate railway having developed a

well-ordered rate schedule, they have all met and agreed upon a unified

scheme; which as far as possible harmonizes all conflicting interests.

The gradation of rates rising with increasing distance from New York

on each separate road, is adjusted to the corresponding gradation of

rates of its neighbors on either side. The result is a series of rate

zones, lying more or less concentrically about the terminal point.

These zones are highly irregular in width and area, but possess one

feature in common. Each remoter zone is one stage higher in rates than

its predecessor. This relationship is indicated by the cross section

diagram herewith. This cross section, of course, differs from the

diagrams heretofore shown. It is purely geographical, being taken, not

along one single railway but as the crow flies--straight across the

whole trunk line territory. But in order to appreciate the significance

of this elaborate scheme, one should imagine a whole series of such

progressively rising rates, radiating out along the different lines

of railway. Connecting the corresponding levels or stages upon each

one with those of its neighbors, the concentric zones are immediately

outlined. The advantage of such a broad scheme is that it generalizes

the single line tariff; taking into view every place, no matter how

small and irrespective of its location whether upon a through line or

merely a local transverse one. Every town, no matter how insignificant,

is assigned a place in a logically evolved plan. Such would seem to

be the ideal of rate construction, toward which all traffic managers

should strive.

* * * * *

The foregoing description of the development of a mileage tariff is

applicable to only a part of the traffic. A very large volume of

tonnage,--said to be not less than seventy-five per cent. in America,

sixty-three per cent. in Prussia and fifty per cent. in the United

Kingdom,--moves under special rates made in quite another way in

response to the exigencies of commercial competition. The making of

these freight rates in practice is an extremely complicated matter.

No single road is independent of rates made by its rivals--rates

applicable not only to competing commodities and markets, but also as

affected by apparently the most remote and disconnected contingencies.

Thus railway rates, as has well been said, are not a set of independent

threads; they form a fabric. They are so interwoven everywhere that if

one thread be shortened, it will cause a kink in the fabric that may

run almost anywhere. In order to understand this it will be necessary

to describe somewhat in detail the nature of competition as applied

to transportation; and then to show by a few concrete illustrations,

the various factors which actively enter into the determination of

specific rates. Laymen and legislators do not sufficiently appreciate

the extremely delicate nature of the work. Much discussion relative

to railway competition seems to be based upon the assumption that it

consists in the main of the competition of railway lines more or less

parallel or else operating under substantially like conditions. As a

matter of fact competition in transportation is to a large degree far

more complex.

Railway competition is of three entirely distinct sorts. These may

be denominated, respectively, competition of routes, competition

in facilities and competition of markets.[76] The first of these,

competition of routes, as the name suggests, is limited to the

activities of the carriers alone. It occurs whenever two railways

are exposed to identical commercial conditions both at the point of

origin and of destination. The rivalry is direct and physical. The only

competition possible is that concerning the route by which traffic

may move between those two points. Such competition is most likely to

arise between more or less parallel lines, as for instance between the

various trunk roads from New York to Chicago. The classic instances

in our history are of the rate wars due to the West Shore and the

Nickel Plate, which were built for the express purpose of engendering

competition with the then existing lines,--the New York Central and the

Lake Shore, respectively. The same sort of simple competition prevails,

of course, between a railway and a parallel canal or other waterway,

as, for instance, between the Erie canal and the trunk lines, or the

Illinois Central and the Mississippi river. Such simple competition as

this, where confined to railways alone, almost inevitably leads to one

of two results: the roads may remain independent, preventing ruinous

rate wars by pooling; or else, as a result of long continued cut-throat

competition, the bankrupt road may be bought up and merged with the

solvent one. This was the fate of the old New York and New England

railway, finally purchased by the New Haven system; of the West Shore

and Nickel Plate lines; and of the Kansas Pacific, unloaded on the

Union Pacific by Jay Gould. The nature of railway competition is indeed

such that no other result than consolidation or pooling can ensue.

Weyl is right in his observation that,--"Strictly speaking, permanent

competition can exist, not between railroads struggling for the same

traffic; but solely between those railroads which have no territory in

common."

This first form of competition of routes or, as it has been called, of

alternative routes, often obtains where conditions of competition are

more obscure than in these simple instances above named. In the rivalry

for the imported plate glass or crockery traffic between the trunk

lines and the Gulf roads, the competition is none the less of routes

between Liverpool and Chicago, although the water carriage by way of

New Orleans or Galveston is so much more roundabout. Freight actually

moves from Boston to Chicago by a line 1786 miles long, via Asheville,

N. C., while the direct distance is only 1004 miles.[77] From St. Louis

to Meridian, Miss., is 512 miles by direct rail line; yet traffic

may move over 2000 miles going to New York and then around.[78] The

map on p. 271, showing the various rail and water lines concerned in

traffic between New York and the little town of Troy, Ala., shows

how widespread are the ramifications of competition of this sort.

Manifold instances of such roundabout carriage have been elsewhere

described in full.[79] They differ from the competition of parallel

routes, however, in the important regard that absorption of the long

lines by the short ones becomes both physically and financially

impossible. Whenever a large area like the Pacific slope is devoid

of manufactures, and wherever the source of supplies is sufficiently

concentrated, as, for instance, in the manufacture of agricultural

implements which are almost exclusively made in or about Chicago, we

still have to do with a clear case of competition of routes, although

a great number of carriers may participate in the business. When

molasses or rice are only to be had from New Orleans, the centre of

such business, the carriers to all tributary consuming points compete

for the routing of it over their own respective lines. These carriers

may operate either by land or sea or by a combination of both; and

they may transport commodities by the most roundabout ways.[80] The

determinant feature, however, distinguishing this class of competition

is neither the mode or carriage nor its length; but is found in the

fact that the commercial conditions at both ends of the line, points

of origin and destination, are identical for each participant in the

business. Direct competition of routes, therefore, has to do with pure

transportation,--the creation of place values,--and this being the

case, the relative cost of service is always a factor of moment.

Competition of facilities, the second of the three phases of railway

competition above mentioned, deals, as its name implies, not at

all with the rates charged but with the facilities or conveniences

afforded. Such competition is confined solely to rivalry for

business at the established rates. Immediately on the appearance

of any departure from these conditions the question becomes one

of competition of either of the other two sorts. An instance of

competition of facilities would be the introduction of reclining

chairs or of a superior service in passenger business. When the Rock

Island system offered such facilities without an extra charge, it

became necessary at once for others to meet this competition in the

same way that they would meet a reduction of rates. Any reduction in

time of transit for freight business between two given points without

extra charge, would in the same manner give rise to competition

of facilities. Such facilities, however, as might have a distinct

money value, as, for instance, free storage, cartage, demurrage or

milling-in-transit, any one of which practically amounts to giving

value without charge, are, of course, equivalent to a reduction of the

rate; and do not belong in this class of considerations at all. Only

those conveniences or facilities, which, while attempting to secure

business may not be compounded for money, should be classified in this

group. It should also be observed that competition of facilities may as

readily arise between parts of the same railway system or under pooling

agreements to maintain rates, as between distinct and independent

companies.[81] And such competition between parent and child often

arises. Thus, for instance, business was as actively solicited as ever

by the Pennsylvania and the Baltimore & Ohio in competition during the

several years of financial control of one by the other prior to 1907.

The New Haven railway may compete with its own water lines around Cape

Cod or on Long Island Sound. But in all of these instances the cardinal

feature to note is that the competition is always at the established

rate. For New England, although the New Haven system and the Boston

and Maine do not compete on rates at their points of contact, there

is constant rivalry in respect of facilities or service. The same

thing is undoubtedly true of the Atchison and the Southern Pacific

in the carriage of California fruits. Although operated under pooling

agreements, yet they were competitors in the matter of the service

offered. Each sought an enlarged volume of tonnage, but not by cutting

the agreed rate.

The third form of competition in transportation is dependent upon

the competition of markets; and is not in reality direct competition

between carriers at all. This is the most difficult of all forms to

understand.[82] It is certainly in many cases more than a "euphemism

for railway policy."[83] Yet although indirect and often obscure,

it is of fundamental and conclusive importance in the determination

of freight rates. Commercial competition deals not with a mere

choice of routes, but with alternative markets. The carriers act,

not independently and of their own volition, but only as agents or

representatives for their constituents, the shippers. They may become

tools or weapons in the hands of merchants or manufacturers who are

the real contestants. It is largely in this sense that it is so often

alleged, and rightfully, that railway traffic managers oftentimes do

not _make_ rates at all. Their energies are bent to the analysis of

those circumstances by which their rates are made for them.

The production or preparation of commodities for final consumption

falls naturally into two distinct parts; the creation of form value,

succeeded by the conferring of place value. Transportation is concerned

alone with the latter process. Of these two operations, the latter,

the creation of place values, is by far the more elastic and adaptable

process. The grower, the miner or the manufacturer has his first costs

more or less rigidly fixed by natural or human conditions; such as the

fertility of the soil, the grade of ore, the prevailing scale of wages,

and so on. His proximity to the status of a marginal producer depends

upon his relative position in these respects. With the carrier, matters

are more contingent. Including within its reach, as it does, many

grades of producers and consumers, each more or less rigidly held bound

by his own circumstances and conditions, as above said, the carrier is

able to exercise a wide range of choice in fixing that margin of value

created which it reserves for itself. And at all times, by reason of

the factors set forth elsewhere, primarily its subjection to the law

of increasing returns, this intermediate share of the carrier tends to

adjust or accommodate itself to the end that it may discover or produce

a wider margin between values in the hands of producer and consumer,

respectively. This may be best accomplished by a progressive widening

of its field of activities, that is to say, by an enlargement of its

physical reach and scope. It is always striving to lower the cost of

production made by the marginal producer. Its motto must ever be, to

get more business, if not right at home by search for it abroad--and

this always with the chance that the greater the distance between the

producer and the consumer, the greater the possible margin of place

value remaining as its individual share.

This ever-present incentive to widen the market carries with it a

direct consequence. A market is a commercial area characterized by a

prevalent equality of prices. Phenomenal development in this respect is

characteristic of the United States. For many commodities the market

is coextensive with the national domain. It is the chosen function of

transportation agents, by rail and water, to ensure this result; to

preserve an equality of prices, despite the variety of producing and

consuming conditions. The railway is the agent by which the market is

thus widened and rivalries are thus equalized. In railway parlance

this is what is known as "keeping everyone in business." The following

quotation from the Senate Committee Hearings of 1905 adequately

describes the process: "I am interested in the erection of a mill

that has just been completed, and sometime since I was figuring on

the question of a smokestack. I wanted to have that stack built out

of brick that is burned in New Jersey, and that is several hundred

miles away. It is a long way to ship freight from New Jersey to North

Carolina. A quotation was made me by the stack builder, whose office

is in New York, and I remarked to him, 'That price is prohibitive; I

cannot pay that price for that stack.' He said, 'That is the best I

can do; but if you will tell me what you can afford to pay for that

stack, in competition with home-burned brick, I will see what I can do

with the railway people.' He said, 'All right; I will take it up with

the railway people.' His quotation included the delivery of the brick

and the erection of the stack at my plant. It would require something

like about fifty carloads of brick to build that stack. Within a week

he had his price revised, and gave me a satisfactory quotation and

took my contract for the stack. Of course he had to get a special rate

from the railway people, because there is no regular tariff on brick

from New Jersey to North Carolina." In this instance the railways

actually created this new business by so adjusting the margin between

the minimum cost of making brick in New York and in North Carolina,

as to make it possible for the traffic to move. The special rate here

mentioned, however, should be carefully distinguished from a secret

rebate offered to one contractor as against another in the same place.

This commodity rate, while special to meet a particular contingency,

was open to any other shipper similarly circumstanced. The student

cannot too carefully discriminate between these two sorts of special

rates. They are constantly confused in the public mind. The effect of

these open commodity rates, is not to create difference of opportunity

between individuals, but to generalize economic conditions and equalize

prices throughout wide areas.

The most satisfactory way to describe commercial competition as applied

to carriers is by concrete illustrations. There are two distinct

varieties or degrees of it, which may be denominated primary and

secondary. These might as properly, perhaps, be called simple and

complex, or direct and indirect. Of these, the first concerns those

cases wherein a commodity undergoes no physical transformation between

producer and consumer. Shipments are usually direct. Only one rate is

involved. Shall St. Louis and the South, for example, be supplied with

salt from the Kansas or Michigan fields?[84] This is a case of pure

transportation,--the creation of place value, alone. The Aroostook

farmers of Maine compete in prices with the potato growers of Michigan

in the New York market. Each district is usually represented by a

railway, dependent upon the prosperity of its particular constituency.

Competition of markets is usually more keen where a number of

carriers are concerned, each representing its own clients; but it

may conceivably arise as between several markets served by the same

company, especially with the growth of great railway systems. The

Southern Pacific must insure a rate from California on oranges to

eastern markets, as compared with the rates over the southern roads

from Florida, sufficiently low to warrant the venture of capital in the

industry.[85] Marble from the quarries of Vermont and North Carolina,

and paving blocks from the Lithonia district in Georgia and from

Wisconsin or South Dakota, must meet in Chicago on even terms. Such

competition, although simple and direct, recognizes no national bounds.

Copper from Montana must be laid down in Liverpool at rates to permit

of meeting the price on Chili bars from South America. Our entire grain

and cotton crops must be transported at rates which will enable them

to hold their own in European markets. The California raisin has, in

this manner, had to make its way into Eastern markets in the United

States against the pressure of importations from Spain, as described

in another place.[86] The cotton mills in New England and in the

South must have their output carried to China under conditions which

will enable them to meet the price made by the British manufacturer.

This last instance, however, introduces us to the second form of

competition; inasmuch as a double transportation is involved first from

the fields to the mill, and thereafter from the mill to the consumer.

Up to this point it has appeared as if, in making distinction between