Olive Oil Prices and Rents

Even though none of the estates had any olives, some of their rents do show striking correlations with oil prices. The results for Group I estates are set out in figure 6.4 and table 6.6. Figure 6.4 makes it quite clear that, with the exception of the decade between 280 and 270 B.C. , oil prices and rents for Group I estates were tightly linked. Indeed, 74 percent of the variation in rents can be explained simply by variation in oil prices (table 6.6). While we must be cautious about the implications of the single datum point available for before 290, there can be no doubt that from about 270 B.C. on, changes in the price of oil predict the changes in rent levels for these estates very well.

The story becomes more complicated in Group II estates. Figure 6.5 appears to reflect relations between oil prices and rents very like those iden-

[Full Size]

Figure 6.4.
Correlation of Olive Oil Prices with Group I Estate Rents, 304-179 B.C.

tified for Group I estates; very interesting again, but difficult to interpret because of the extreme paucity of data points, are the apparent correlations between high prices and high rents (304–303 B.C. ) and the rise in rents in the decade 280–270 while oil prices fell. For the decades after 270 B.C. ,

[Full Size]

Figure 6.5.
Correlation of Olive Oil Prices with Group II Estate Rents, 304-179 B.C.

however, the subgroups of Group II, although showing at least some correlation with oil prices, diverge from one another and from Group I in a very interesting way: the degree of correlation between oil prices and rents varies inversely according to the number of vines (table 6.7). Group IIA estates, which depended heavily on wine production, show poor correlation with oil prices. Only 23 percent of rent variation can be accounted for by changes in oil prices, the significance of the correlation is poor, and the amount of auto-correlation as measured by the Durbin-Watson is unacceptable. An attempt to correct for the latter eliminates all significance from the relationship.^[16] In other words, decisions about the level of rents renters were willing to pay for Group IIA estates were essentially independent of oil-price levels. Rents for Group IIB and IIC estates, on the other hand (table 6.7 B, C), moved in conjunction with oil prices. Half of the variation in Group IIB estates, with a moderate number of vines, is linked to oil-price change; and fully 66 percent of the variation in Group IIC can be attributed to oil-price changes. This last figure approaches the result for Group I estates.

The first two patterns apparent for Group I estates seem reasonable.

High oil and high estate prices might be linked before 290 B.C. , as many scholars have thought (but see below), but the data points are so few that it is very difficult to say. That rents and oil prices should move in different directions in the 270s occasions no surprise, since the estates had no olives and therefore their rents should not have been affected by oil prices.

But the results for the years after roughly 270 B.C. are striking. The pattern virtually reproduces the results for barley: the fewer the vines, the greater the link between oil prices and rents. These results seem strikingly to confirm D. Rathbone's view that "agricultural prices in general, including that of olive oil, are likely to have followed similar price trends in the long term."^[17] But the linkage between oil and grain prices implied by the correlation of both with rents is itself puzzling. No clear mechanism accounts for it. The natural history of the two crops is very different (grain was planted in the fall/winter and harvested in the spring/summer; olives were harvested in the fall); their demands on labor fall at different times of year and at different levels; their rainfall requirements are strikingly different as regards both amount and timing. Most important, they cannot have been substitute goods, like wheat and barley. When wheat was scarce, consumers turned to barley, and the rising demand drove the price up. But it is hard to see why consumers unable to get barley would turn to oil, especially as oil was a relatively expensive product in antiquity anyway.

This correlation presents many questions, and it is impossible to offer a definitive solution. Reflection on the local situation on Delos and the neighboring Kyklades, however, may suggest a tentative hypothesis. In chapter 5, I argued that, after a period of dependence on extra-local sources for olive oil in the late fourth and early third centuries, Delians and their immediate neighbors planted olives and began to supply local needs from local sources. (The failure of the hieropoioi to plant olives on Apollo's estates must be attributed to reluctance to make the long-term investment on rented property.) This development reduced and stabilized oil prices. It would also have had an effect on the local price of land, for once the olives began to produce, their yields would have become part of the income-producing capital of the property they grew on. Oil prices would then generally have become a factor in the equation by which land values, and so rents, were set. This linkage would be still stronger if it was customary in the Kyklades—as would seem very likely—not to plant olives in dense orchards but to scatter them about the landscape with enough space in between to grow grains. Such a mechanism could explain why even rents for

Apollo's estates, which had no olives, followed oil prices. The power of this explanation must not be exaggerated. Unlike the correlation between rents and barley prices, that between rents and olive oil prices breaks down on occasion. This indicates that the tie is not direct, as I have argued for barley, but indirect, perhaps through the mechanism suggested here. Indeed, it seems likely that, rather than trying to explain rents as a dependent function of oil prices, we should seek a larger explanation to encompass both. My suggestion that planting olives helped to stabilize land prices is such a larger explanation, but, as we shall see in chapter 7, there are still other, more general, factors that need to be considered.