Thursday, January 13, 2011

Railways paper: six panels up

We now have six 'panels' each representing an estate's rent for the period 1832-1899. That is a total of 337 observations allowing for missing data and the 'instruments' needed to adjust for temporal effects. Malcolm just sent me the track records for the last estate, Badminton....I held my breath while the computer did the work (takes a few minutes). I was blue in the face when the answer came out---still significant! So we have a model that looks like this:

Family home for the owners of the Thorndon Estate
Note the positive sign for track, which means that an increase in railway within 40 km of the estate increases the rent. The effect isn't large, probably because the difference in costs between using the railway and 'droving' the livestock along a road wasn't a large part of the total farm budget. BUT it is statistically significant, and that is what counts.

One of the six estates is Thorndon, pictured here. Think of the heating bill! No wonder they had to raise the rents of all their tenant farmers!

Next I am going to work on  a separate regression for Holkham Hall, a big estate in Norfolk for which Mi has been getting me the data. Holkham is one of the six, so we already have some knowledge of this estate. It is in Norfolk, on the east coast, about a hundred miles north of London.

No comments: