Transport, urbanization and economic development in England and Wales c.1670-1911
Coal prices and trading costs
One central aim of our Leverhulme project is to estimate transport costs between towns and parishes across Britain. Towards that end, GIS files of market towns and the transport network were created at various dates from 1600 to 1911. We also collected estimates of freight rate rates per ton mile, travel speeds per mile, and distances traveled per day from the literature.[i] In a preliminary analysis, we combined the aforementioned data to estimate transport costs by inland waterway or coast between the 100 largest towns in 1801. Specifically we used the network analyst tool in ArcGIS to identify the lowest transport cost between any two towns using the network and cost parameters. The following map shows for each town, the average water freight rate in shillings per ton to all other towns in 1830. As can be seen, average transport costs for inland towns on waterways were similar or only slightly larger than towns on the coast. While preliminary, the estimates have significant implications in assessing the market access of towns at a key moment during the process of industrialization.
While our aim in the Leverhulme was to analyze all transportation modes, the project was primarily focused on travel by roads, rivers, canals and railways, reflecting the areas in which had prior expertise. We now want to develop new areas of the project in relation to coastal shipping. As many historians have noted, the coastal trade was extensive, and made up the majority of ton miles in the important coal trade up to the late nineteenth century.[ii] Unfortunately, there is limited information on coastal freight rates before 1850 because most cargoes were carried on the ship owner's account with no freight rate being charged.[iii] Several works in the literature have proposed using coal prices to infer transport and related costs as an alternative to having direct observations on freight rates.[iv] The idea is that coal is produced in sharply delimited regions, and thus its price in non-supplying locations must equal the supplying location's price plus 'trade costs'. The trade costs include duties levied at ports, freight costs paid to ship owners, insurance costs, and potentially others. In the case of the London coal trade, Hausman finds that duties represent just under half of the price difference between coal in Newcastle and London in most years during the eighteenth century.[v] The implication is that 'shipping costs' (i.e. freight costs, insurance, etc.) could be as much as one-half the price difference. We have also investigated some advertised coastal freight rates between Newcastle and London in the early 1840s, and found them to be around 40% of the price difference between Newcastle and London coal.
Our future plan is to estimate total trade costs using the difference in coal prices across markets in Britain, and ideally with information on duties at individual ports we will estimate the share of trade costs associated with duties and the share under the general category of shipping costs. This method requires obtaining coal prices in a sample of markets across Britain. Thus far, we have collected data on coal prices in 55 markets from 1691 to 1703, on 30 markets in 1785, 480 markets in 1843, and 90 markets in 1913.[vi] We are also exploring the use of gas company records, poor law accounts and the account books of coal merchants, and institutional and household purchasers of coal to enlarge our sample of coal prices between the late seventeenth and early nineteenth centuries.[vii] We have done some preliminary investigations to establish the feasibility of using these sources.
We are aware of the issue that coal comes in quality varieties, and they were priced differently within the same market. In some cases, like London in the early nineteenth century, we know the price for different varieties of coal and we can examine how the reported London market price accords with the quality distribution[viii]. From these cases, we think we can develop reasonable assumptions about how market price observations generally relate to quality.
In a significant extension to previous work, we propose to use estimates of coal trade costs to infer coastal shipping costs per km. The methodology involves several steps outlined in the following. First, we identify all coal supplying regions. The main ones in the eighteenth century were in the North-east, South Wales, Lancashire and Cheshire, Yorkshire, East Midlands and West Midlands.[ix] Second, we identify all customs ports where supplying regions shipped their coal by coastal vessel. The main coal suppling ports were Newcastle, Swansea, and Liverpool. Third, we use our GIS coastal network to determine the shortest distance between all ports and one of the coal supplying ports. For example, Newcastle, and its surrounding ports, were the closest suppliers on the east and southeast coast. Swansea, and the surrounding ports in South Wales, were closer to the southwest coastal ports.
The fourth step is to estimate the relationship between the coal price in each port and its assumed supplier. An example is shown below for a sample of east coast ports stretching from Newcastle to Chichester in the south. The horizontal axis shows the distance to Newcastle. The vertical access shows the coal price in pence per ton in constant 1840 prices.[x] The data show a clear positive relationship between distance to Newcastle and the coal price in 1700, 1785, and 1840. A crucial parameter is the estimated slope. It shows the increment in the coal price for every extra KM of distance on the coastal network. The slope also gives an estimate of the coastal trade cost per ton KM, provided that coastal transport is cheaper than inland waterway and road, which by all indications it was. Following from this assumption, we estimate that in 1700 the coastal trade cost was 0.81 pence per ton KM (measured in constant 1842 prices). It fell to 0.41 pence per ton KM in 1785 and 0.28 pence in 1843. Using information on duties we can then adjust the trade costs per km and infer shipping costs. If our Newton Trust application is successful, we plan to implement this methodology on a larger scale and estimate average shipping costs by coastal ship across Britain, and possibly by coastal region as well.
We plan to go one step further and use coal price differences in inland areas to estimate shipping costs by inland waterways. We discovered in the course of the Leverhulme project that shipping costs by river and canal could differ in notable ways. As one example, Maw has estimated that canal freight rates were 15 to 36% of road rates between Manchester and Liverpool. But between Manchester and Hull they were 60% (most likely because of the heavy lockage required to pass over the Pennines, pp. 75-76).[xi] We think we can systematically measure the differences in freight rates by river and canal using coal price differences. These inferences require more assumptions than previous steps. We begin by identifying all inland markets that were connected to a port by an inland waterway. Before the canal era, it is safe to assume that these inland markets obtained their coal by coast and then by their connecting inland waterway. In the canal era, it is more complicated and we need to be incorporate the possibility of supply by inland coalfields like those near Birmingham, West Yorkshire, and Bristol. In our simplified setting where the origin of the coal is coastal, we would calculate the distance from the inland market to the port by its waterway. The estimated trade cost per KM would be the price difference between port and inland market divided the distance by inland waterway.
An example is the inland market of Guildford. It was connected by the river Wey to London from the mid 1600s.[xii] According to our current sources, the average coal price in Guildford in 1700 was 286 pence per ton. In London, the average price was 258 pence per ton. The difference of 28 pence or 46.7 in constant 1840 prices would represent the transport cost by inland waterway. As Guildford was around 50 KM from London by the river Wey, the estimated trade cost per ton per KM by inland waterway would be around 0.93 pence in constant 1840 prices. Notice the similarity between the estimated coastal and inland waterway shipping rates in 1700. We would expect coastal shipping costs per KM to be much lower than inland barge freight costs because of the scale effects of larger coastal vessels. However, this does not seem to be the case for trade costs per KM. We think the similarity was due to the duties on coal in London which could represent half of the total trade cost.
If our Newton Trust application is successful, we plan to develop this methodology further using our team's expertise in network modelling. A more rigorous approach is to identify the least cost network path between an inland market (say Guildford) and its supplying port (say Newcastle) by coast, by inland waterway, by road, or some combination of the three assuming a given value for the per KM trading cost by each mode. The assumed values and least cost paths will imply a total trade cost which can be compared with the actual trade costs. Conducting the same calculation for many combinations of possible modal trade costs and markets will give a set of predicted trade costs and actual trade costs measured by coal price differences. We would then select the trade cost parameters for coast and inland waterway that minimizes a statistical loss function. Donaldson's work on Indian railways has successfully employed a similar method to infer trade costs from differences in salt prices.[xiii] One member of the team has worked with Donaldson before and could expand on his approach.
To summarize, our plan is to collect a large sample of coal prices across markets in benchmark years. From these data we can infer trade costs in the market for coal. With additional data on duties we can also infer the share of trade costs associated with transportation and related shipping costs. Finally, we plan to use our data on transportation networks to infer coastal, inland waterway, and possibly road shipping costs per km. The estimate will ultimately feed into an analysis of market access across British towns during the industrial revolution.
[i] Observations of freight rates and travel speeds are available in various publications. The following is an abbreviated list: Jackman, William T., The Development of Transport in Modern England, (1962); Hawke, Gary Richard, Railways and economic growth in England and Wales, 1840-1870, 1970; Gerhold, Dorian. Carriers and Coachmasters: Trade and Travel before the Turnpikes, 2005; Peter Maw, Transport and the Industrial City. Manchester and the Canal Age, (2014).
[ii] See Armstrong, John. The Vital Spark: The British Coastal Trade, 1700-1930, (2009), for estimates of the size of the coastal trade. Note there were some fluctuations in coal shipped by coastal trade in the nineteenth century, and coastal coal shipments were regularly matched by rail only on the eve of the First World War.
[iii] See Armstrong, the Vital Spark, and Ville, Simon, " Defending productivity growth in the English coal trade during the eighteenth and nineteenth centuries," The Economic History Review (1987), for discussion of freight rates in coastal shipping.
[iv] See Hausman, William J, "The English coastal coal trade, 1691‐1910: how rapid was productivity growth?" The Economic History Review (1987); Harley, C. Knick. "Ocean freight rates and productivity, 1740–1913: the primacy of mechanical invention reaffirmed," The Journal of Economic History (1988).
[v] Hausman, William J, "The English coastal coal trade."
[vi] The prices for 1693 to 1700 come from John Houghton's weekly reports on Husbandry and trade and are printed in Rogers, James E. Thorold. A history of agriculture and prices in England: from the year after the Oxford parliament (1259) to the commencement of the continental war (1793). Vol. 4, (1987). For 1785 prices see Report of the commissioners appointed to inquire into the several matters relating to coal in the United Kingdom. Vol. I. General report and twenty-two sub-reports, BPP 1871. For 1843 prices see Account of the Prices of Articles of Consumption at the Poor Law Unions, BPP 1843. For 1913 prices we use Cost of living of the working classes. Report of an enquiry by the Board of Trade into working-class rents and retail prices, together with the rates of wages in certain occupations in industrial towns of the United Kingdom in 1912, BPP 1913.
[vii] For examples of price data and sources for the pre-statistical period see Hatcher, J., The History of the British Coal Industry: Before 1700: Towards the Age of Coal (1993), pp.557-89.
[viii] On the heterogeneity of coal and coal prices see: Mitchell, B.R., Economic Development of the British Coal Industry 1800-1914 (1984) pp. pp. 263-82 and Church, Roy., The History of the British Coal Industry, Vol. 3, 1830-1913 (1986), pp. 48-70.
[ix] There were further small coal fields in Cumberland, North Wales, and the South-west:. Church, Roy., The History of the British Coal Industry, Vol. 3, 1830-191, pp ix, 3.
[x] The prices are adjusted using Greg Clark's consumer price index, but note they are not sensitive to the index used.
[xi] Maw, Transport and the Industrial City.
[xii] See Willan, Thomas Stuart. River navigation in England, 1600-1750. Psychology Press, (1964).
[xiii] Donaldson, Dave. Railroads of the Raj: Estimating the impact of transportation infrastructure. No. w16487. National Bureau of Economic Research, (2010).