Focus for the 'supply response' paper

I have only just been able to work out a suitable 'motivation' for our third paper, British agriculture from about 1870 onwards. We've been collecting all the data, county by county, but with no fixed purpose. I think I have it now: regional differences in response to the American 'grain invasion' of the late 1870s. In that period, the American railroads opened up and importing wheat became competitive. The price of grain fell dramatically and many farmers were ruined. The brighter farmers switched to livestock and used the cheaper grain as their input, making good profits. Not all regions adapted in the same way. Why? That is what we are going to find out.

Some railway stats found

Mi and I have been looking for statistics of agricultural commodities carried by rail, just got a bit lucky. I checked the 19th Century House of Commons Parliamentary Papers under 'railway returns'. And found very detailed statistics for all sorts of carriage. Cattle are lumped in with other goods etc until about 1860. Then we get detailed record by head of cattle, pigs, etc. We now have to work out how much effort is required to extract the data and whether the end result will be worthwhile.

Neural Network Results

I am trying out some software which builds a neural network from a dataset. It looks for connections---like our own brains. I used the data Malcolm has been working on for three counties. The results look like this:

Variable Impact Analysis     wheatyieldn100 47.3206%     mktdist 25.1332%     barleyyieln100 8.3316%     oatsyieldn166 8.3279%     elevation (m) 5.4441%     mktpop 4.9268%     grassland 0.4124%     arable 0.1035%
It is a surprise to me that distance from the market town is so important---I'll run the test again for different counties.

The two papers we`re working on

The task has been to write three papers for peer-review and publication in academic journals. So far one paper (on the politics behind some aspects of Peel`s ministry of 1841) is in its final draft stage. The other two are:
agricultural rents in the southwest of England in 1836; and changes in agriculture in nine counties as a result of changes in transport of agricultural commodities by the `new`railway systems.

For the rents paper, Malcolm is getting data from the 1836 Tithe Commission into a format for statistical analysis. So far we have three counties done: Devon, Wiltshire and Gloucestershire. There are two more to go: Dorset and Herefordshire. I have noticed that the relationship between arable rent and wheat yield is very strong, as one would expect. More interesting is the relationship between arable rent and distance to market town. According to von Thunen`s theories of location, the greater the distance, the lower the rent. This holds for Devon, which is far to the west. It doesn`t hold so well for counties more to the east, or closer to larger cities such as London. The distance to market seems to become less important as we approach large centres of population. This implies more integration into the national economy.

To test this hypothesis, Malcolm is creating buffers around six market towns in a line (roughly) from west to east. In each buffer we will have fifty data points (or parishes). If the hypothesis holds, then the significance of distance to market will decrease as we move to the east.

 Map above shows Devon with the 96 datapoints (parishes) marked. Devon was relatively isolated from the rest of the British economy as a result of its geography. Moving east from Devon towards London, proximity to market town diminishes in importance for the calculation of rents.

Mi is building up a large dataset of crops and prices for nine counties in England for a period of approximately twenty years. The counties are: Bedford; Berkshire; Buckinghamshire; Cambridgeshire; Cheshire; Cornwall; Cumberland; Derby; and Essex. Soon, Malcolm will begin to build a database of construction of railways in those counties. We will prepare an animation which shows how the amount of each crop produced in each county changed over time as the railway altered the costs of bringing goods to market. This will give us some insights as to how we might make a formal statistical test of the relationship between crop production and transportation,