• Tidak ada hasil yang ditemukan

Chapter 1 Introduction

2.4 Result 1: Prices to Africans and Competitive PricesPrices

2.4.1 Data

To test this hypothesis, I use newly-collected data on prices in Africa, prices in France, and transport costs. I focus on nine main commodities exported by French Africa:

peanuts, palm kernels and oil, cotton, cocoa, coffee, rubber, timber, and bananas.

The commodities included in the dataset account for 80% of the value of all exports from West and Equatorial Africa during the whole colonial period.

Prices in Africa and Quantities Exported

Colonial customs statistics reported the total quantity and value for each exported commodity from each colony every year. These statistics were registered at the local customs offices and then aggregated at the colony level. The reported values were usually official market prices in Africa (valeurs mercuriales), established by decree by the General Governor of the colonies on the basis of reports of local commissions of evaluation. After World War II, the reported evaluations were often values of the commodities at the loading port, including transportation costs to the port and exit taxes.8

Table 2.1

Components of Prices: Cocoa, 1958–59

Togo Congo Cote d’Ivoire

price to producers 100 82 90

transport to port 4 12 5

taxes, insurances, stockage 32 30 31

price at African port 136 124 126

Source: elaboration fromDocuments et statistiques—Ministre de la France d’Outre-mer, Service de statistique, 1958–59. Prices are in current francs per kg.

Using these customs statistics, I collected data on prices in Africa and quantities exported from each colony for nine main commodities between 1898 and 1959. I exploited numerous yearly issues of different colonial publications, including, but not limited to, statistical reports of the Ministry of Colonies, Bulletins Economiques of the various colonies, and Annuaire Statiques of West and Equatorial Africa.9

Given the variety of the sources and the length of the period considered, the names of the territorial units for which the custom statistics are reported change over time and sometimes data are reported only for larger territorial units. To solve these issues, I first tracked the variation in the names of colonies. Then, I assigned each

8These values are a good proxy for the actual prices to producers: detailed data on prices to farmers, inland transport cost, and tariffs from the late 1950s show that, after controlling for trans- portations costs, differences in prices at African ports are almost exclusively due to differences in prices to producers. See Table 2.1.

9See the appendix for more details on the sources.

commodity from a larger territorial unit to the smallest territory (colony or group of colonies) that I could identify by excluding those colonies which do not produce that specific commodity.10 I deflated all prices in 1900 French francs.11

Prices in France

I collected prices in France from various issues of the Statistiques Mensuelles du Commerce Ext´erieur de la France, a monthly publication by the Direction G´en´erale des Douanes reporting the total values and quantities of the commodities imported from the French colonies in every year. As a control, I also used different issues of the Annuaire Statistiques de France reporting similar information. I deflated all prices in 1900 French francs.

Not all exports from French Africa went to France. Nevertheless, given the impor- tance of the French market, using export prices in France is a good approximation.

By 1949, France was the destination of about 80% of the total exports originating from its African colonies (Duignan and Gahan, 1975). Moreover, French prices are highly correlated with world market prices, as shown in appendix 2.8.1.

Transport costs

Since extensive data on transportation costs between Africa and France are not available, I constructed estimates of shipping costs for each colony-commodity-year in my dataset according to the following procedure. First, I computed the distance to Marseilles from the closest African port for each colony.12 Then, I used data on average freight rates from the West African coast to France for the main exports in 1938 to compute the average shipping cost per km for each commodity in 1938.13 Finally, I multiplied this measure by the distance to Marseilles for each colony (both

10For example, all cocoa exports 1898–1907 recorded as from French Congo (including Gabon and Congo) are assigned to Gabon because there was no registered cocoa production in Congo before 1927. Territorial units involved are AEF, French Congo, AOF, and Senegal-Haut Senegal-Niger.

11Inflation data come from France-Inflation.com (2013).

12The main ports are identified from the map reported at page 149 of Duignan and Gahan (1975).

The distance to Marseille is computed by using http://ports.com/sea-route.

13Documents et statistiques—Minist`ere de la France d’Outremer, Service de statistique, 1949–52.

West and Equatorial Africa) and by an index of transportation costs between 1898 and 1959 with base 1938=1 from Mohammed and Williamson (2004).14 In the empirical analysis, I will consider also inland transport costs, by controlling by the distance of each colony to the coast.

Overall, the dataset includes 1717 observations (colony-commodity-year) and I have data on prices in Africa, prices in France, and transportation costs for 1466 of them. The years covered are: 1898–1914, 1920–1951, 1953, 1956, 1958, and 1959.

Table 2.2 reports the summary statistics.

Table 2.2 Summary Statistics

mean st.dev. min max N

price in Africa 0.71 1.18 0.01 17.9 1714 transport cost 0.04 0.02 0.02 0.06 1717 price in France 1.28 2.34 0.06 19.63 1469

In 1900 French francs.

Dalam dokumen Extractive Institutions in Colonial Africa (Halaman 31-34)