Entrepreneur Reports as Evidence of Profit Expectations

The profit expectations of the first transcontinental entrepreneurs and investors cannot be identified by inferences on the level of wilderness of a region, allocation of subsidies, or ex-post bond transactions. Entrepreneurs’ reports, composed of preliminary survey reports and stock and bond prospectuses, are an alternative source of evidence for profit expectations.

Preliminary survey reports usually contained an engineering survey and market research. The market research first identified the observed market equilibrium and then predicted the equilibrium expected once the railroad entered. Operation costs were estimated and operation profits deduced. Entry into the transport market was determined by the ratio of operation profits to construction costs. Entrepreneurs used these reports to communicate formally with investors. As the basis of stock and bond prospectuses, the reports thus contained the key public information used by investors to make purchase decisions.Footnote ⁶

The entrepreneurs’ reports may be used to identify profit expectations in three different ways. First, these reports include statements by the entrepreneurs explicitly indicating whether they expected the railroad to be profitable. Such statements are not generally to be taken at face value. In the case of the first transcontinental, entrepreneurs were motivated to overstate profit expectations to capital markets and to overstate social benefits and understate profits to Congress. Since it is not possible to infer the overall sign or magnitude of these effects, these statements are reported without additional informational value attached.

Second, performance and publication of a preliminary survey, or of its results in a stock or bond prospectus, are indications of expected profits. A preliminary survey was a sunk and expensive activity, costing at least $30,000 for a railroad such as the transcontinental, or more than twice the annual sales of the average 1850s American manufacturing establishment (BL, Judah 1857; Williams 1988, pp. 36–37; Davis 1855, p. 24; Census 1865). Like any R&D investment, performance and publication of a preliminary survey or a prospectus reflect profit expectations.

Third, the reports explain the method used in the 1850s to evaluate a railroad investment opportunity and produce a profit estimate. Examination of a sample of preliminary surveys, stock and bond prospectuses for other 1850s American and English railroads, and of American and English engineering textbooks reveals that the method used by the first transcontinental railroad entrepreneurs is identical to that of other successful railroad entrepreneurs of the 1850s.Footnote ⁷ The method may then be described as standard in the 1850s for investment appraisals by engineers, entrepreneurs, and investors alike. The intuition behind this method makes it possible to develop the structure of a simulation model that estimates expected profit. The parameters of the model are estimated with information publicly available before the subsidies were granted.

The model of expected profit outcomes uses only information available ex-ante for the investment decision and has two advantages. First, the model provides a counterfactual setting to examine the ex-ante private investment decision where incentives to build the railroad depend exclusively on operation profits. It abstracts “experimentally” from the investment decision the effects of corrupt schemes, insider's profits, subsidies, or capital market distortions. Second, the model allows for testing the sensitivity of the expected profit estimates to specific data or assumptions.

Used in this way, the entrepreneurs’ reports provide evidence on the profit expected by the average first transcontinental entrepreneur and investor. The main caveat of the estimate is that it does not use private information that may have been available to privileged investors at the time. This is not crucial when the purpose of the exercise is to determine whether an average entrepreneur or investor should have expected profit.

Profit Expectations as Stated by Entrepreneurs

Initially proposed as a single-stage project (1845–1859) by at least six different groups of entrepreneurs, the plans divided the project into two stages after the Nevada gold rush and Judah's survey of the new route. The first stage with an expected construction cost of $13.3 million was half the cost of the comparable Corps route. The price and traffic of wagon transport across the Sierra Nevada made the observed revenue $5.7 million per annum. Entrepreneurs promised to set the rail route's prices below the prices by wagon and estimated traffic and operation cost, and all predicted expected profit of $3.7 million per annum, or 27 percent of construction cost. Information in Judah's preliminary 1861 survey shows careful and detailed consideration of the first stage as an investment opportunity.Footnote ⁸

The preliminary survey also contained superficial information on the second stage. A detailed survey for the second stage was completed by the Central Pacific and the Union Pacific between 1862 and 1866, while the first stage was being built. Surveying improved the route and reduced the expected construction cost by 33–50 percent compared with the $86.7 million predicted by the Corps. The price and traffic of sailing ships and Panama steamers to and from China and California made observed revenue $31 million per annum. Entrepreneurs promised to set the rail route's price in between the total price along the routes for the sail ships and the Panama steamers, and they expected rapid traffic growth. After estimating the operation cost, the predicted expected operation profit was $27.6 million per annum, or 31 percent of the construction cost (BL, Judah 1861, 1862a; Union Pacific Railroad Museum (UPRM), Durant 1864; British Library Manuscript Collection (BLM), Dodge 1867; BL, Fisk and Hatch 1867; UPRM, Cisco 1868).

A Model of Formation of Profit Expectations

To formulate the structure of the simulation model, the 1850s method of investment decisions is recreated in preference to a more sophisticated modern approach. The emphasis here is on how entrepreneurs and investors gauged their profit expectations at the time.

The 1850s and modern approaches share the same intuition and the same structure. Construction costs are estimated with detailed engineering studies. Expected revenues are estimated by identifying the observed market equilibrium on the alternative transport mode, then assuming/deducing a pricing policy, a price elasticity of demand, and a growth rate of transport demand. Subsequently, the expected market size and share for the new railroad are estimated. Expected operation costs are estimated by adapting the costs of railroads already running to local geographical conditions. Finally, operation profits are deduced and compared to construction costs. Note the similarity between this approach and what is called a profit function in modern economics. Also note that the observed market equilibrium plays a key role anchoring the analysis firmly circa 1862. The changes to this equilibrium are relatively static; they involve movements along the same demand schedule. Construction of the railroad plays no role in inducing future migration, and therefore no feedback effect on future transport demand is incorporated.

However, the 1850s and the modern approaches also differ. First, the 1850s documents are not always explicit about the details underlying their analysis, such as the functional form of the demand function, discounting methods, or the calculation of the promised price. Second, simpler methods were used in the 1850s to estimate the values of the model parameters. In contrast, assumptions made to carry out this analysis are highlighted in this section.

The simulation model develops a scenario in which entrepreneurs motivated only by operation profit face the decision to enter the transport market. Two assumptions are necessary.

First, entrepreneurs may buy the right of way for a fixed fee L and without going to Congress. The assumption excludes from the analysis the perverse incentives that entrepreneurs faced when requesting the right of way from Congress. By assuming it is possible to buy the land for a fixed fee without going to Congress it is then possible to focus on operation profits, excluding rents from subsidies or corruption. The value for L is set at 1 percent of construction cost, based on western land prices and the share of land costs in construction cost for eastern railroads.

Second, entrepreneurs face an elastic capital supply at interest rate r. The assumption abstracts from the distortions the Civil War imposed on investment decisions circa 1862. The purpose is to consider if the project could have been profitable under “normal” capital market circumstances. The assumption also excludes the possibility that capital markets were more likely to fail in financing the transcontinental venture than in financing other railroads. The point is not to argue that the capital market of the 1850s did not fail but to highlight that, once the project was divided into two stages, each stage was not more uncertain or larger than earlier projects funded by the domestic and international capital markets.

The technical uncertainty of construction was within the range of uncertainty exhibited by existing railroads. Examination of the ex-ante and ex-post evidence on the construction process of the first transcontinental indicates that, contrary to what many believe, the railroad did not face unprecedented technical challenges. In 1855 the Corps stated that the route was practicable, entrepreneurs improved the Corps route in 1859–1862, and actual construction of the railroad shows that construction proceeded roughly as planned, as measured by cost, technical specifications and time. Crucially, no deviation from standard construction practices was expected, and the actual construction experience gives no indication that construction techniques were improved. The expected construction costs of each of the two stages of the first transcontinental were within the range of construction costs of other railroads and canals built in the 1850s and funded in the capital market. The best indication that the domestic and international capital markets were large enough to fund the transcontinental is that the transcontinental and the Suez Canal were actually financed almost exclusively with stock and debt issues (and no current taxes), while competing for funding with the Civil War and other European investment projects. The discount rate adopted for the simulation model (9 percent) is higher than the highest yield (7 percent) offered by infrastructure projects during the period to raise private capital through long-term debt.

The average 1850s entrepreneur or investor would have begun by identifying the railroad's expected construction cost based on the route chosen between origin location i and destination j. The expected total construction cost, C, is the sum of the right of way fixed fee, L, and the flow of construction cost, c, discounted at the discount rate r

The next step in the investment decision is to estimate expected revenues. Let N be the number of submarkets in which the railroad is expected to operate. Each submarket is defined by a pair of origin i and destination j locations. As the entrepreneur divided the project into two stages, he identified five submarkets. In the two submarkets of the first stage, the rail route competed with wagon roads across the Sierra Nevada. In the three submarkets of the second stage the rail route competed with sail ships around Cape Horn and ships to Panama, a transfer to the railroad across Panama, and a transfer to ships thereafter.

After the submarkets have been defined, the observed equilibrium on the alternative transport mode in a relevant submarket ij is identified as q_ij tons (or passengers) per annum and P_ij dollars per ton (or passenger) trip. For instance, after performing a traffic survey, Judah estimated “the yearly average of freight [on the wagon road over the Sierra Nevada] to be 120 tons per day at an average price of six cents per pound” (BL, Judah 1862a, p. 51).

Next, the entrepreneur considered how entry of the rail route would have diverted existing transport demand in submarket ij and then predicted the expected price and traffic. In the first stage, Judah argued that “a reduction in [transport] price … and a saving in [transport] time … would give satisfaction to merchants and consumers, and secure every pound of Washoe freight over your road” and predicted the price, traffic, and receipts (BL, Judah 1862a, pp. 52–54). The rail route was expected to have a technological and cost advantage compared to wagon roads. The entrepreneur figured that the highest price he could charge would have been the sum of the observed price paid for the existing transport, P_ij, and the non-transport trade costs saved by switching to the railroad, B_ij. If he set an expected price, lower than (P_ij + B_ij), the new railroad would capture 100 percent of the traffic. After setting expected price, the entrepreneurs then predicted expected traffic,

The preliminary survey reports and stock and bond prospectuses identify the observed and expected equilibrium but do not explain how the 1850s entrepreneurs calculated the expected price and traffic. Since the nature of the changes the entrepreneurs considered were generally movements along a fixed demand schedule as the price changed, presumably the observed and the expected quantity of traffic were on the same demand schedule. An expected demand function that gives all possible pairs of expected price and traffic is defined as

where parameter a_ij gives the sensitivity of expected traffic to expected price and h_ij is the demand intercept when the transport price is zero and reflects the economic size and other relevant origin-destination pair-specific effects on trade. The quantity demanded from the expected demand function must equal the observed traffic when evaluated at the observed price, hence

In the second stage, the rail route was expected to offer the China and California trades and passengers greater safety, a drier and more stable climate than routes that cross the tropics and “destroy all animal and vegetable products [and] injure all manufacturing goods,” and reductions in transport time (GK, Whitney 1849, pp. 17, 27, 36; GK, Degrand 1849, p. 11; McDougall 1854; UPRM, Cisco 1868, pp. 22–25). Goods were expected to arrive at their destination in better shape and goods not traded before, like fresh fruits, could now be shipped on the railroad. Preliminary survey reports and bond prospectuses indicate that the entrepreneur expected to set a price “half the cost of the steamship [Panama] route … and shortness of [transport] time decides … most cargoes will find their natural transit over [the transcontinental] railroad … excepting some very heavy or bulky articles of comparatively low values” (UPRM, Cisco 1868, pp. 23–24). That is, the entrepreneur expected to set the price between that of the Panama route (high) and sail around Cape Horn (low), diverting a substantial part of market share from sail and displacing the Panama route.Footnote ⁹

Following the intuition that changes introduced by the transcontinental route were of a static nature and therefore the observed and the expected equilibrium quantity are both on the same demand schedule, the expected demand function is defined as

The sensitivity of expected traffic to expected price is given by parameter b_ij. The origin-destination pair-specific trade effects are captured by u_ij, which is set to allow the expected demand function to equal the observed total traffic via the Cape Horn and the Panama routes when evaluated at the observed low price.

Once expected revenues have been calculated, expected operation costs are estimated. The entrepreneur used eastern railroads’ observed costs to derive the constant average cost per passenger or per ton transported, o, and the expected operation cost function .Footnote ¹⁰

To select the optimal expected price (and implied traffic), it is assumed that the entrepreneurs chose the price maximizing operation profits. The Credit Mobilier scandal revealed that entrepreneurs were profit seekers and that they might have maximized personal profit. However, the relevant variable to determine whether investors should invest, and whether government should intervene, is the first transcontinental's total operation profit, η, defined as the sum of operation profits over all relevant submarkets for each stage. The average entrepreneur's assumed optimal expected operation profit in each submarket was identical for each period t of operation.

The entrepreneur would then compare expected total construction cost, C, with the flow of expected operation profits, η, and derived the entry condition given by η/C. In a more modern context, it is important to explicitly account for a discount rate, and the entry condition is given by

where T is the total lifetime of the project. For investors to decide to build the transcontinental, the sum of the discounted stream of total operation profits must be higher than the sum of the expected stream of construction costs.

Decision to Build the First Stage

The expected construction cost for the first stage is $13.3 million, spread evenly over five years (all values are in 1860 dollars). The cost of purchasing the right-of-way is 1 percent of construction cost, $133,000, the discount rate is 9 percent, and the project lifetime is 25 years. Revenues from transportation start in the sixth year and continue until the project's life ends. The transportation market between Sacramento and Virginia Station, Nevada featured two submarkets: freight and passengers. The observed traffic and price by 1862 was 43,800 tons per year and $120 per ton per trip, and 13,505 passengers per year and $30 per passenger per trip, while the non-trade transport cost B_ij is zero. The expected demand function non-price term is calibrated using the observed traffic and price. The average operation cost was $1.83 per ton per trip and $1.36 per passenger per trip. Recall that the first stage was expected to compete on prices below the price ceiling set by wagon roads. See sources and details on parameter values in the Data Appendix. The baseline scenario equilibrium is calculated for a_ij, implying a price elasticity of demand between 0 and −1. The profit-maximizing solution to the model is trivial. The entrepreneur should set a price for the transcontinental such that the expected price is just marginally lower than the observed price. Baseline freight profits should have been expected to be $5.2 million per annum and passenger profits $387,000 in line 1 of Table 1.

Once the baseline expected operation profit for the first stage is calculated, it is possible to compute the entry condition. The ratio of operation profits per annum to construction cost, the 1850s entry condition, is 39 percent, while the contemporary investment threshold was 15 percent. The expected net present value (NPV) of the project is $24.5 million. The average entrepreneur and investor should have expected the first stage of the first transcontinental to be profitable.

The robustness of the baseline scenario result is examined by calculating the maximum change to each single parameter required to make the expected NPV equal to 0. Actual construction cost was $14.1 million, implying a small cost overrun, but it could even treble and the project would still be profitable in line 2 of Table 1. The NPV of the project would have been positive even if the observed price and traffic declined rapidly and permanently by 40 percent, the discount rate almost trebled, the project's life shortened to less than half, or the revenues were delayed by almost a decade. The average entrepreneur or investor would have perceived the first stage of the first transcontinental to be profitable.

Decision to Build the Second Stage

The second stage of the transcontinental, from Virginia Station, Nevada, to Omaha, Nebraska (including a segment Sacramento-Oakland), completed the full route to the Pacific Ocean. Because the first stage was already operating profitably when the decision to build the second was to be taken, it was known that the technically more difficult first stage had been successfully built and at a lower construction cost than forecast by the Corps. Entrepreneurs had completed the surveys for the second stage and were now ready to build it.

Table 1

BASELINE SCENARIO AND SENSITIVITY ANALYSIS – FIRST STAGE

* Operation cost of freight service X 43.6 and operation cost of passenger service X 21.9.

Notes:

Price and quantity are the profit maximizing choices. Lines 2, 3, 4, 5, 6, and 7 are results from changes in parameters calculated to make NPV = 0.

The second-stage project's life starts in year six of the project and ends in year twenty-five. The expected construction cost is $57.2 million, spread evenly over years six through ten. The right-of-way fixed fee and discount rate are the same as for the first stage. Three submarkets comprised the transport market: freight in both directions between California and the eastern United States (California trade), freight in both directions between China and the eastern United States (China trade), and passengers in both directions between California and the eastern United States (California passenger traffic). Revenues start flowing in year 11 and continue until the end of the project's life. Recall the second stage was expected to provide new good attributes and the price was to be set at a level between that of sail ship around Cape Horn and that of the Panama route.

Based on 1850s information, the total traffic in the California freight submarket was 221,088 tons per year with a high price of $140 per ton per trip and a low price of $16.83 per ton per trip. The values for the China freight submarket were 119,774, $140, and $17.49, respectively, and for the California passenger submarket the values were 66,153 passengers per year, $252.60 per passenger per trip, and $50 per passenger trip. The first transcontinental route also involved the eastern rail network between Omaha and the final destination in the eastern United States, at $20.50 per ton per trip and $14.96 per passenger per trip, while the China trade submarket included the Shanghai and California segment, at $6.83 per ton per trip. The expected operation cost is the same as for the first stage. See sources and details on parameter values in the Data Appendix.

The baseline scenario results indicate that entrepreneurs and investors should have expected profits from all three submarkets in line 1 of Table 2. The profit-maximizing freight price would have been approximately $70 with traffic of 135,000 tons and the passenger price would have been $128 with 40,000 passengers per annum. Annual operation profit for the California trade would have been $4.3 million, $2 million for the China trade, and $4.4 million for the California passenger submarket. The project's NPV would have been $7.8 million. Expected total operation profit as a percentage of expected construction cost, the contemporary entry condition, is 18.6 percent, higher than the 1850s 15 percent threshold. The sum of the NPV for the first and second stages combined would have been $32.3 million. The average entrepreneur and investor should have expected the second stage and full first transcontinental to be profitable.

Note that the expected NPV of the second stage less the PV of the expected construction cost of the first stage gives a lower bound estimate of expected profits of the single stage project, as it was planned between 1845 and 1859. The average entrepreneur or investor should have expected $3 million in losses from the single-stage project. This implies that, as planned in the 1850s, a railroad with no revenues from local traffic should not have been expected to be profitable.

The robustness of the baseline scenario result is examined by calculating the maximum change to each single parameter required to make the expected NPV equal to 0. The NPV remains positive even if the expected construction cost of $57.2 million rose by 24.3 percent in line 2 of Table 2. Recall that the actual construction cost was only $50.5 million. The NPV is still positive even after the high freight price fell by 18 percent, or the high passenger price fell by 57 percent.

Table 2

BASELINE SCENARIO AND SENSITIVITY ANALYSIS – SECOND STAGE

Sources:

Price and quantity are the profit-maximizing choices. Lines 2, 3, and 4 are results from changes in parameters calculated to make NPV = 0.

The results presented for the baseline scenario are based on conservative criteria. The profits would have been still higher than the baseline if the railroad could have practiced price discrimination, if there had been input substitution, if there were economies of scale and density, if new trade was created, or new sources of revenue from Colorado's gold rush in 1859 or the opening of the Japan trade were considered.Footnote ¹¹