Bancroftiana, Number 123 Fall 2003: Towards Estimating the Demand for California Wine: 1870

Towards Estimating the Demand for California Wine: 1870–1920

Economists define the demand for a good as the quantity of a good one is willing to buy at alternative prices, holding one’s income, tastes, and the price of related goods constant. To estimate an empirical demand model for California wine, we need actual data on wine quantities, wine prices, income of consumers, the price of substitutes or complements for California wine, and a proxy for tastes. In addition, since the price of wine and the quantity of wine are determined simultaneously, we have two unknowns in a single equation, i.e., we cannot know price without knowing quantity, and vice versa. This is called the identification problem. To solve it, we need to have instruments. Instruments are exogenous random shocks that are correlated with the price of wine, but are not otherwise related to the quantity demanded.

I became interested in estimating a demand model for California wine for the latter half of the 19th century because of the availability of two instruments: phylloxera and the sharpshooter. Phylloxera is a root louse that erodes the productivity of certain vines. Phylloxera was first observed in California at Sonoma in 1873. Other things remaining equal, the effect of phylloxera is to reduce the supply of grapes and thus raise the price of wine. The sharpshooter eats succulent plant tissue; thus, it destroys a vine quickly. The sharpshooter struck Southern California vineyards in 1884, and by the mid 1890s it had destroyed almost 25,000 acres of vines. From the 1860s through the early 1880s, Southern California had produced more wine grapes than Northern California.

The San Francisco Merchant, Especially Devoted to the Encouragements of the Productions, Manufactures, Commerce of the Pacific Coast, Vol. XI, Number 12, December 28, 1883. This journal is replete with advertisements for rooted vines and cuttings, wine agents and merchants, and related business endeavors.

On my first trip to The Bancroft Library last September, I learned that having good instruments was not, by itself, a sufficient condition for completing the research. While I was checking in, the receptionist asked what I was searching for, and I told her California consumer wine prices from 1850 to 1900. She found only three sources of wine prices, and all were from hotel restaurant wine lists.

I learned that week that almost all California wine from the 1850s through 1906 was made, aged, blended, and bottled in San Francisco near the Embarcadero. Grapes or wine were shipped by boat or, later, rail from vineyards or wineries to the city. With the exception of Inglenook, wineries did not bottle their wine at their winery until 1944. At that time, the defense department took over all rail and water transportation capacity in the Bay Area to fight World War II. For the most part, before 1944 consumers in the city or the wine country brought jugs to the wineries and their jugs were filled. The upshot of this crude distribution system to the consumer is a dearth of price data for estimating a demand model for California wine.

California has shipped the majority of its wine production out of state since the 1860s, and price and quantity data are available on shipments by sea and rail; thus, wholesale prices are readily available. Starting in 1890, the Pacific Wine & Spirits Review (PWSR) began publishing retail prices of California and French wines in San Francisco outlets. Articles in PWSR explain the relationship between wholesale and retail prices so we can construct a consistent set of prices from 1890 to 1920. Published reports from the California State Board of Agriculture provide production levels of wine from 1850 to 1923.

In 1906, the San Francisco earthquake and resultant fires that lasted three days destroyed most of the large wineries in the city. In addition, 25 million gallons of wine were lost. The earthquake is clearly another instrument that will help us estimate the demand curve. In addition to price, our demand curve will include the price of substitutes (French wine) and income changes in New York, New Orleans, and California, respectively.

By the end of 2003, I will have preliminary estimates of the determinants of the demand for California wine. I would like to thank the staff at Bancroft for their help and guidance.

—Douglas Brown