LAND RENTAL VALUES AND THE AGRARIAN ECONOMY: ENGLAND AND WALES, 1500-1912 Gregory Clark, University of California, Davis, CA 95616 ([email protected]) February, 2001 I use the rents and prices of land held by charities in England and Wales to estimate statistically the nominal and real rental value of farmland from 1500 to 1912. Real land rental values increased fourfold between 1500 and 1912. Combining these estimates with farm wages and rates of return I calculate agricultural output from factor payments, as well as output per worker, from 1500 to 1912. Finally the productivity of agriculture is calculated. These new series suggest that measured agricultural productivity doubled between 1500-39 and 1860-69. But the productivity growth was fairly evenly spread over this long period, and was much less in the years 1760 to 1860 than standard accounts of the Industrial Revolution such as Crafts (1985) assume. Thus these estimates imply a substantial reduction in estimates of output growth in England in the Industrial Revolution era. Further the rate of productivity growth in England in the years 1500-1789 is no greater than the growth rates Philip Hoffman finds for the Paris Basin. Finally, contrary to expectation, the source of productivity growth before 1869 is overwhelmingly growing yields as opposed to growth of labor productivity. Introduction The rental value of farmland in England in the years 1500 to 1914 is not just a matter of antiquarian interest. For this value is the key to understanding when productivity growth took place in English agriculture in the years 1500 to 1914. In combination with information on wages, capital returns and product prices we can use rents to measure agricultural progress. This paper employs a statistical method to measure farmland rental values which is explained and tested below. Using the derived rental values the implications of the new rent series, in combination with series on wages, returns on capital and prices, for the agrarian history of England is then explored. Determining the rental value of farmland is not easy, since in early years much farmland was not rented for its current rental value. Instead land was held on a bewildering variety of tenures – customary leases well below market values, leases for lives where the current rent has little relation to current market conditions, renewable leases with low annual rents but large entry fines and so on. This paper seeks to abstract from the real conditions of tenures to estimate for the years 1500 to 1912 what land in general would rent for if offered for rent in a free market. Estimates are derived for land of constant characteristics for four regions in England and for Wales. To adjust to actual rental values allowance is made for the extent of common land in each period. The series is then compared to existing rent series. It proves to be similar to some, but to diverge sharply from the recent series offered by Turner, Beckett and Afton (1997). Sources The source of rental values for this paper is land held by charitable trusts. Here I give a brief description of the sources. A much more detailed analysis is given in Clark (1998a). Over the course of the eighteenth and nineteenth centuries the British Parliament conducted a number of investigations of the activities of charities which documented the known history of the land they held, and its current rental value. The earliest set of reports, the Gilbert Returns, show the 2 responses to a request sent to local parishes for information on the current incomes of charities in 1786. This was followed by the mammoth Charity Commission or Brougham Commission of 1818-1837. This inquiry published 32 reports containing 26,987 pages, which give not just the current state of charity assets, but also often a history of how and when the land was acquired. The 28,880 endowments for charity reported on held 442,915 acres of land, nearly 2% of all farmland. A permanent Charity Commission was established in 1853. This published a digest of the current income from land and securities of all charities in the years 1868 to 1875. It also issued annual reports, which contain some information about the property of charities whose affairs were brought to the attention of the commission. In 1884 the Commission issued a report on the City of London Livery Companies which gave details of the current leases of their property, and the state of the property circa 1860. Finally in 1889 the Commission launched a parish by parish update of the Brougham Commission inquiry. This new inquiry lasted till 1913, but succeeded in covering only Berkshire, Devon, Durham, Lancashire, London, Wiltshire, and the West Riding of Yorkshire in England and about half of Wales. These later inquiries often contain information about land values and returns on capital even in the years before 1837. Thus the charity inquiries can give information on the value of land before 1786, in 1786, in the years 1818-37, in the years 1862-75, and in the years 1889-1912. In addition to the various Charity Commission sources, the rent of this charity land is sometimes recorded in other sources. The reports on the Charities for Elementary Education printed in the Parliamentary Papers in 1906 and 1908 give the current rents of land held by charity schools in Cornwall, Kent, Northampton and Stafford. The volumes of the Victoria County Histories devoted to individual parishes give details of the current rents of charity lands, or of the 3 sale year and price at sale. Those counties where at least some of the parish volumes were completed before 1914 thus give information on current land rents in the years 1905-1914. The problem with the 1786 Gilbert Returns, the 1860s Digest, the reports of the Charities for Elementary Education, and the Victoria County Histories is that they give only the current rental value of land, which may have been leased some time before. But the two sets of detailed Charity Commission reports confirm that in the course of the nineteenth century there was a large-scale move towards shorter leases of charity lands. Thus in the years 1818 to 1829 65 percent of land was on yearly leases, but by the 1890s this had increased to 85 percent. Consequently for the later nineteenth century the current rent of land will generally indicate the current market rental value of the land. From these reports I have extracted a subset of observations where an estimate of the rental value of land can be made from the rent on a newly formed lease, or the price of land at purchase, or the current rental value, or just the current rent paid where there are indications this rent was effectively paid on a yearly tenancy without any formal lease. These observations on the rental values of charity land have been formed into a data set containing 32,202 observations on 19,475 land holdings. Many pieces of land are observed more than once. The location by parish of most of the plots is given, as is the land area. There is also a variety of information on the physical characteristics of the land available for many of the plots: land use, associated farm buildings, enclosure status, enclosure history, land tax status, tithe status. Table 1 shows the information available on plots observed more than once with at least 5 acres per house, in total and by five major regions. There are 19,696 such linked observations drawn from 7,236 farms and plots of land, of which 7,733 observations are of plots or farms of 20 acres or more. The average plot size is 47.8 acres in this sample of linked observations. This 4 Table 1: Linked Observations on Land Rental Values by Decade Period All All holdings North Midlands South East South West Wales 20+ acres 1500-39 25 16 - 8 14 3 - 1540-59 35 22 4 18 9 4 - 1560-79 27 18 - 11 14 2 - 1580-99 45 22 3 6 31 5 - 1600-09 34 19 2 13 14 4 1 1610-19 77 35 11 19 36 9 2 1620-29 120 56 7 30 60 22 1 1630-39 114 56 17 29 35 31 2 1640-49 91 47 7 20 43 20 1 1650-59 123 57 15 42 37 24 5 1660-69 92 38 16 24 29 22 1 1670-79 171 73 34 46 56 30 5 1680-89 185 61 24 71 34 52 4 1690-99 211 91 31 58 55 61 6 1700-09 205 101 29 63 64 45 4 1710-19 235 108 44 65 59 66 1 1720-29 288 132 47 93 59 79 10 1730-39 320 125 63 117 65 61 14 1740-49 250 86 59 68 43 60 20 1750-59 222 83 47 67 36 54 18 1760-69 214 84 56 60 37 54 7 1770-79 212 89 47 59 43 55 8 1780-89 550 167 73 224 122 116 15 1790-99 312 148 40 78 91 94 9 1800-09 780 380 117 247 197 196 23 1810-19 1,879 847 333 476 392 633 45 1820-29 4,099 1,535 1,118 1,288 920 718 55 1830-39 2,573 979 108 875 932 429 229 1840-49 97 45 33 10 21 8 25 1850-59 222 112 90 45 31 14 42 1860-69 1,648 584 254 513 521 311 49 1870-79 1,179 438 353 265 363 77 121 1880-89 368 175 109 35 67 39 118 1890-99 1,245 482 824 52 88 86 195 1900-09 1,144 340 200 373 105 464 2 1910-12 304 82 11 86 56 151 - all 19,696 7,733 4,226 5,554 4,799 4,099 1,038 5 Sources: Land Values data set. Notes: The earliest period 1500-39 includes two observations for the south east for the 1480s included in hopes of extending the data as much as possible. The areas are composed as follows: North - Cheshire, Cumberland, Durham, Lancashire, Northumberland, Westmorland, Yorkshire. Midlands - Bedford, Berkshire, Buckingham, Derby, Huntingdon, Leicester, Lincoln, Northampton, Nottingham, Oxford, Rutland, Stafford, Warwick South East - Cambridge, Essex, Hampshire, Hertford, Kent, Middlesex, Norfolk, Suffolk, Surrey, Sussex. South West - Cornwall, Devon, Dorset, Gloucester, Hereford, Monmouth, Shropshire, Somerset, Wiltshire, Worcester. 6 is not much less than the average occupancy size of 60.5 acres for England and Wales in 1888, though since the charity land was often leased by persons owned or leased other land we cannot directly compare these averages. The regions are the North, Midlands, South-East, South-West and Wales. These regions were drawn more for convenience than for any underlying presumption about geology or climate. But they do have differences in climate that make the south-east the predominantly arable area of the country, with only 26% of land in permanent pasture in the South-East in 1866, compared to 51% in the North. Thus rent trends may differ across the regions. The charity sample of land is not representative of land as a whole. Aside from a relatively small average plot size charity land is also found more where there were more people. Thus it also over samples densely populated parishes. Table 2 shows, for example, the distribution of the land area of England across parishes of different population density in 1841 versus the distribution of charity land observations. While only 7% of all land was in parishes with more than 10 people per acre in 1841, 14% of the rent observations from the charity data were in such parishes. Thus if rent trends in the more densely populated parishes in 1841 were different than for the less densely populated parishes the charity data will be misleading about national trends. A national rent index is calculated from the charity data by using statistical methods to correct for the ways in which charity land is unrepresentative in terms of regional spread, amount of common land, plot size, and distribution across parishes of different population densities. The calculation is done in two stages. First a benchmark is established for the level of rents nationally and by region in the years 1820-24 using the broad cross section of rent data available for charity estates in the years 1818-1837. Average levels of rents for each county in England 7 Table 2: Share of Parishes by Population Density, England Population Land area (m. Share of Land Share of all charity Share of all Density (people acres) Area rent observations large plot per 10 acres, observations 1841) 0-1 4.77 0.15 0.04 0.06 1-2 9.94 0.31 0.20 0.25 2-3 7.67 0.24 0.28 0.28 3-4 3.32 0.10 0.15 0.15 4-5 1.61 0.05 0.07 0.06 5-10 2.40 0.08 0.13 0.09 10+ 2.10 0.07 0.14 0.10 All 31.81 1.00 1.00 1.00 Sources: Parish Land Areas and Population Densities, Parliamentary Papers 1852-3a, 1852-3b. 8 and Wales are estimated. In the next stage the sub sample of plots observed more than once over time are used to estimate indices of average rents for each of four regions of England and for Wales for the years 1500-1912, which are then adjusted to the 1820-24 benchmark. Tithe Throughout I estimate rents inclusive of the tithe paid by the tenant, since this is what matters for use of these series to estimate agricultural productivity or the relative rental of land compared to wages. But in only about one sixth of the cases do I know whether land paid tithe or not. And where the tenant paid tithe again in only about one sixth of cases do I know the amount of the tithe paid. For land which paid tithe but the amount paid is unknown I estimate the amount of tithe paid using the information from the cases where we know both tithe payments and rent. The tithe as a proportion of the total rental value of the holding (including tithe) varied by region. Table 3 shows the estimated share of the total rental that the tithe represented for each of the four regions of England and also Wales for the years before and after 1840. I split up the data in this way because the Tithe Reform Act in 1836 changed the way tithe was levied which might affect the future tithe burden. Since tithe was in principle 10 percent of gross output, and rent and tithe together was only about 40 percent of gross output in the nineteenth century, the tithe should have been about 25 percent of the sum of rent and tithe. On average I find that tithe was a much smaller share of the sum of rent and tithe than this, being only 12.2 percent before 1840 and 14.8 percent after. Tithe as a share of rent was lowest in the north, being there only about 7 percent of the sum of rent and tithe. It was highest in the southeast where it averaged about 19 percent of the sum of rent and tithe. The adjustment to rents in cases where tithe was known to be paid by the tenant 9 Table 3: Tithe as a Share of the Total Rental Value of Land (in Percent) Region Pre-1840 Pre-1840 Pre-1840 Post-1840 Post 1840 Post-1840 Observations Mean Standard Observations Mean Standard Error Error North 23 6.5 1.09 175 7.7 0.49 Midlands 49 11.5 1.09 97 16.5 1.03 South East 23 15.9 2.64 27 21.7 2.24 South West 9 14.9 2.14 67 13.4 1.14 Wales - - - 36 10.6 1.47 England 104 12.2 0.93 402 14.8 0.69 10