On a whim, I don't know why, I typed "Inca economics" into Google (okay, I was procrastinating). After poking around some, I came across this. The purpose of the paper is to measure network characteristics of the Inca empire and test hypotheses about organization, but I've focused on the parts of the larger effort describing the economic relationships. I included one of the figures to help visualize the descriptions. So, for anyone who might also feel like reading about Inka economics, in particular the network relationships that were involved and how they may have been used to maintain political power, here's some of the paper [Update: I wonder if there is a debate in ethnohistory over whether to use Inca or Inka similar to the debate in economic orthography over whether to use heteroscedasticity or (the correct) heteroskedasticity?]:
A Network Analysis of Inka Roads, Administrative Centers, and Storage Facilities, by David Jenkins, University of Arizona, Ethnohistory 48.4 (2001) 655-687: ... Staple Finance and Wealth Finance The Inka in the early fifteenth century were a chiefdom or perhaps an anomalous early state (Bauer 1992) of about twenty thousand people with a fairly simple social organization based on kinship ties and ruling hereditary chiefs. Initially their territory was limited, centered on what would become the city of Cuzco. Over the course of a hundred years, from about 1430 until the Spanish arrived in 1532, the Inka dramatically expanded their empire, incorporating by political maneuvering and outright conquest some eighty distinct polities into the Inka state. These conquered groups included other expansive empires, such as the highly socially stratified Chimu on the north coast, as well as small-scale states, chiefdoms, tribes, and autonomous communities scattered throughout the highlands.1
As part of their strategy of conquest, the Inka forcibly resettled many people, for example, the 135,000 to 270,000 Wanka who lived in the Xauxa region of the Peruvian highlands and who were subjugated around 1460 (D’Altroy and Hastorf 1984). Some resettled people, called mitmaq, became a source of permanent state labor and engaged in farming, cloth production, metal working, and pottery making. Others remained on their lands but were required to participate in a system of rotating labor assessed at the household level. This rotating labor, called mit’a, involved traveling far from one’s community to farm, mine, weave, build roads, or participate in other state-oriented activities (Murra 1982; Wachtel 1982; LeVine 1987).
Extracting both labor and tribute, the Inka constructed a vast road system centered on Cuzco that connected coastal and highland regions and stretched some thirty-two hundred kilometers from Quito in the north to Santiago in the south. All together there were more than thirty thousand kilometers of Inka roads, which linked twelve million subjects, allowed for their efficient relocation, and proved essential for the maintenance of political rule (Regal 1936; Strube Erdmann 1963; Hyslop 1984; Stehberg and Carvajal 1988). Figure 1 shows Hyslop’s (1984) map of the road system with important Inka sites indicated.
Click on figure for larger version
Two distinct forms of exchange flowed through the Inka road system, both imposed by the state and both disguised in Inka ideology as reciprocal exchange. As systems of exchange, however, they had different structural properties. Staple finance (Polanyi 1968: 186–7) was the infrastructural basis of military expansion and administrative control, while wealth finance provided the means to tie local elites of conquered groups to the Inka political economy ( D’Altroy and Earle 1985). Staple finance resulted in various storage complexes scattered throughout the empire, the contents of which—maize, chicha (maize beer), coca, hot peppers, cloth and clothing, sandals, pottery, weapons, and so on—were used to pay for state activities, including military excursions, elaborate ceremonial displays of Inka beneficence, and administration itself. Huánuco Pampa, for example, with extensive storage facilities totaling as much as 37,100 cubic meters, could support a population engaged in state activities of between twelve and fifteen thousand persons, most of whom were not permanent residents (Morris 1992: 159; Morris and Thompson 1970, 1985: 96).2 Wealth finance, by contrast, was based on easily transportable, high-prestige objects, such as feathers and shells, and resulted in an exchange network that encompassed the entire Inka empire (D’Altroy and Earle 1985).
Staple finance was based on a simple logic. All resources were held by a ruling elite, the Inka. Local access to land, water, and other resources was granted in exchange for local labor, which produced surplus subsistence goods on state-owned lands. Individual communities and individual households within them were allowed to maintain their own productive self-sufficiency, provided that they participated in the mit’a, the system of rotating labor that generated state surpluses (see Murra 1980 , 1982). In this system, local descent groups (ayllus) were obligated to provide each year a quota of adult males from their communities for state projects.
The collection and storage of staple wealth was regionally focused. With a few exceptions, subsistence goods were moved short distances only, usually no farther than a hundred or so kilometers (D’Altroy 1992). One of the exceptions was maize from the Cochabamba valley, which apparently was transported eight hundred kilometers by llama caravan to Cuzco (Wachtel 1982). To accommodate the goods transported to regional centers required the construction of storage complexes and administrative facilities to govern them (Morris 1967). Hatun Xuaxa, for example, had about twenty-seven hundred individual storage units in its vicinity with overall storage capacities estimated at more than 171,000 cubic meters (D’Altroy and Earle 1992). Goods flowed into and were stored at regional centers, such as Hatun Xuaxa, Huánuco Pampa, and Pumpu, but rarely flowed from regional storage center to regional storage center (Morris 1967, 1986, 1992).
The extent of storage astonished the Spanish conquistadors and later chroniclers. Pedro Pizarro (1965 : 191) thought that the supplies were so extensive that they could never in his lifetime be exhausted. Pedro Sancho de la Hoz (1917 [1532–3]: 194–5), one of Francisco Pizarro’s men, noted of the region around Cuzco:
All of the fortress of Sacsayhuaman was a warehouse of arms, clubs, lances, bows, arrows, shields, strong jackets padded with cotton and other arms of diverse types, and clothing for soldiers assembled here from all parts of the land subject to the Lords of Cuzco. . . . [There are] storehouses full of textiles, wool, arms, metal and clothes, and of all things grown and made in their land. There are storehouses where the tributes brought by the vassals to the lords are kept, and there is a house in which are kept more than one hundred thousand dried birds, because clothing is made from their feathers that are of many colors, and there are many houses for that. There are bucklers and shields, beams for roofing, knives, and other tools. [There are] sandals and breastplates to provision the soldiers, in such quantity that reason is inadequate to understand how they could give such a great tribute of so many and diverse things. (Morris 1992: 167–8; his translation)
One conservative estimate of the total volume of storage throughout the empire is one to two million cubic meters (Morris 1967: 155; see also Murra 1980  for a discussion of historical sources on state storage).
Staple finance is fairly common for chiefdom-level societies and early states. It has been found in Mississippian chiefdoms, the Aztec empire, Hawaiian chiefdoms, the Yamota court of sixth-century Japan, and elsewhere (Brumfiel and Earle 1987). It is especially well suited to relatively small, geographically compact polities. When polities based on staple finance are geographically extensive, however, they must be decentralized, simply because of the exorbitant costs of transporting heavy and bulky subsistence goods to a single center. A number of subcenters are required to store and disburse such goods. When decentralized, however, societies based on staple finance are politically vulnerable: Local administrators in charge of an independent source of wealth could easily fund a rebellion.
The Inka used several mechanisms to forestall rebellion. First, the overall administrative structure of the empire was hierarchical: Each regional center answered directly to Cuzco. Each local community similarly answered directly to its regional center. Officials in this hierarchy, often recruited from local elites, had jurisdiction over, in descending order, ten thousand heads of households, five thousand heads of households, one thousand, five hundred, one hundred, fifty, and ten (Means 1931: 292; Rowe 1946: 262; Niles 1987: 53). In return for their labors, officials of these groups of households were granted various rights and privileges, including the expectation of receiving status markers, such as fine cloth appropriate to the official’s position in the hierarchy. As the span of control is quite small, six officials or fewer, the flow of information is reasonably efficient. With this system the Inka intended to control communication and to preclude the exchange of both goods and information between communities and between regional centers, thus emphasizing vertical ties at the expense of horizontal ties (D’Altroy 1992). (There was some variation in the hierarchical structure, especially on the peripheries of the empire, where the number of officials and thus the number of links between them were fewer; see Salomon 1986.)
As a second means to avoid conflict, the Inka forcibly moved rebellious groups to different regions, sometimes a thousand or more kilometers from their homes. The result was a form of internal colonization, which produced in state farms, regional centers, and in the capital at Cuzco a social mosaic in which individual ethnic identities were maintained. Those from the same locations lived together in ethnic settlements, maintaining their languages and styles of dress (Murra 1980 ; Wachtel 1982).
The third and perhaps most important means to avoid conflict was a form of wealth based on relatively lightweight, easy-to-transport, high-prestige goods, which were given to regional and local leaders in exchange for their continued allegiance (Murra 1962, 1980 , 1982; Morris 1967; Wachtel 1982; D’Altroy and Earle 1985; LeVine 1987).
This form of exchange, called wealth finance, also flowed through the Inka road network but transcended regional boundaries and regional centers. It was oriented toward the transportation of status objects, such as Spondylus shells from the north coast, bird feathers from the eastern tropical regions, fine cloth spun from llama and alpaca wool from the central highlands, and gold, silver, and copper from various highland and coastal locations, all of which were sent to Cuzco. Inkan wealth finance was primarily based on mitmaq labor, that is, on the work of permanent state laborers, which resulted in the accumulation of prestige goods in Cuzco, some portion of which was redistributed to local elites throughout the empire (D’Altroy and Earle 1985).
Wealth finance was a means to stabilize the politically unstable system of decentralized staple finance. The use of mitmaq labor, the central storage of status goods, and the centralized system of redistribution combined to bring wealth finance directly under state control. By rewarding local and regional officials with fine cloth, for instance, the Inka intended to tie their economic well-being directly to the Inka state (Murra 1962). In general, however, high-prestige goods, as markers of social status, allowed elites access to subsistence goods but could not be exchanged for subsistence goods (Earle 1987). Staple finance and wealth finance thus circulated in different spheres of exchange.
In broad outline, this is the system I wish to place in a network perspective: two distinct circuits of exchange with different structural properties. One resulted in a massive storage complex with large subcenters and many (perhaps as many as two thousand) intermediate way stations, called tampu. In this system, goods were regionally stable, while people were mobile; the military, royal retinues, mit’a laborers, and so on moved through the road network and had their subsistence needs supplied along the way. The other circuit of exchange was empire-wide and was based on the movement of lightweight, high-prestige goods into and then out of Cuzco. As Murra (1962) and Morris (1986) have suggested, long-distance movement itself probably conferred added value to the already high-prestige goods.
Inka Road Network
... The central location of Cuzco, confirmed by two different centrality measures, is especially noteworthy: It has the highest centrality in terms of the relative efficiency with which it could communicate with any other point in the network and the second highest centrality in terms of its ability to control the flow of information or goods between any two points. Such relative locational advantage gives support to Morris’s (1986: 64) observation about the ideological position of the Inka capital: “One reason [prestige goods] went to Cusco was so that they could come from Cusco.” Cuzco’s high closeness centrality made the movement of goods to and from Cuzco relatively efficient.
Because Cuzco lies on the second highest number of shortest paths between any two points, it is also in an advantageous position to control the flow of information throughout the empire. The famed chasques (Cobo 1964 : 129–31), messengers stationed at regular intervals along the road network, are a spectacular example of this control. Cobo, impressed by the speed with which messages could travel from point to point, noted that orders from Cuzco could be sent to Quito and return to Cuzco in ten or twelve days. Cieza de León (1967 : 70), similarly impressed, noted that “in this way the lords were informed of everything that occurred in their kingdom and dominion and [thereby] determined what appeared to be in their best interests.” Cuzco’s central location facilitated such informational control. ...
Prestige goods, as part of the system of wealth finance, also circulated through the road network. Spondylus shells, honey, tropical wood, birds and bird feathers, fine cloth, gold, silver, copper, and in some instances maize flowed into Cuzco, and some portion was redistributed to regional and local elites (Cobo 1964 : 125)...
Tumbes (point 40), on the Ecuadorian coast, was a significant conduit for Spondylus shells destined for Cuzco (Cobo 1964 : 126). Called mullu in Quechua, Spondylus shells and the objects fashioned from them held great religious significance for the Inka (Murra 1975; Rostworowski 1970, 1975). They appear in many pre-Inka sites in the northern and central Andes and may have entered into central Andean exchange networks as early as 1100 b.c. (Paulsen 1974). Spondylus was not a significant trade item in the preceramic period, however; only a few instances of its preceramic occurrence in the highlands have been documented (Rick 1988: 38).
The habitat of two Pacific species of Spondylus, Spondylus princeps and Spondylus calcifer, range from the Gulf of Guayaguil in Ecuador to the Gulf of California. Both species live in waters twenty-five to sixty meters deep (Marcos 1978). As Murra (1975) notes, millions of Spondylus must have been harvested to satisfy Inka needs. Tumbes, as part of an exchange network stretching into Central America, was at the southern edge of Spondylus habitat. Thus, to ensure that a supply of Spondylus reached Cuzco, the Inka required open lines of exchange from Tumbes to the highlands. In terms of wealth finance, then, Tumbes and the road that linked it with the sierra were particularly important. Cobo (1964 : 126–7) noted that the road from Tumbes to the sierra was one of the two most significant lateral roads linking coast and highland.5 And Francisco de Xérez (1917 : 322–6), one of the 175 European invaders who landed in 1532, observed several storehouses near Tumbes filled with sufficient cloth and food “to maintain themselves for three or four years.” ...
Although Tumbes has relatively low centrality in the Inka road network, its removal would have effectively cut off the flow, or at least much of the flow, of Spondylus to the highlands. As an outer point in the system of wealth finance, Tumbes’s importance was based not on its central location but rather on ... what might be called its strategic marginality.
Other locations with similarly low centrality were important in the system of wealth finance. Turi (point 52), for example, was an Inka administrative center that controlled the flow of precious metals from the Atacama desert to the highlands. ...
...[Staple Finance and Wealth Finance represented] two distinct circuits of exchange correspond to what Earle and D’Altroy (1989), following Hassig (1985: 100–1), call strategies of territorial control and strategies of hegemonic control. They argue that Inka imperial power is best conceptualized as a continuum of strategies, with hegemonic control at one end and territorial control at the other:
With hegemonic control, the core Inka polity ruled indirectly, through clients who retained considerable autonomy in management of local affairs, but were forced to accept military submission, tributary payments in labor and special products, and truncation of economic relations with neighboring ethnic groups. For the core, the advantages of such hegemonic relationships were found in the relatively low costs of control and in the minimal internal security duties.
With territorial control, the core Inka polity extended direct state presence into the dominated regions through a comprehensive restructuring of economic and political relationships, a supervisory administrative presence, and often a series of dramatic cultural changes. Control was high, permitting a high level of extraction, but the degree of intervention was tempered by the high costs of direct administration. (Earle and D’Altroy 1989: 187)
Hegemonic control requires minimal administrative investment. The intent is to produce a surplus from local populations simply by usurping, under threat of attack, a portion of the labor or products that would otherwise have been consumed by that population. By contrast, territorial control requires high administrative investment. Its intent is to tie the welfare of local populations directly to the state. In the Inka empire, territorial control was strongest in the central highlands between Cuzco and Quito, where there were large populations and high agricultural productivity, and weakest along the northern and southern coasts and in the eastern tropical regions. Wealth finance with its centrality correlates would facilitate hegemonic control, whereas staple finance, with its different set of centrality correlates, would facilitate territorial control.
And I thought this section at the end was interesting:
In the Inka road network, constriction points or bottlenecks were frequent. D’Altroy (1992: 86) provides a vivid example of environmental influences on the capacity of one line in the Inka network:
To illustrate the effects of terrain on military movements, Chalcuchima’s army of 35,000 stationed at Hatun Xuaxa at the time of the Spanish arrival may be taken as an example. Single file, about 500 soldiers take up 1 km (cf. Hassig 1988: 65–72). If the porters, soldiers’ wives, and other camp followers composed a group about the size of Chalcuchima’s army, the string would have stretched out to 140 km. If a llama is allowed 3 m in the train, about 333 llamas take up 1 km of road. A llama caravan of, say, 15,000 accompanying the troops would have extended an additional 45 km. The entourage would likely have regrouped, rather than travel single file, of course, and Inka armies were typically deployed in sections, rather than as single bodies. As a unit, however, moving at a rate of 5 km per hour (i.e., 1 league/hour), it would have taken this army and entourage 37 hours to pass through the narrowest point of the road or across a bridge. Traveling constantly from dawn to dusk, the human force would have consumed more than 300,000 kg of food and the llamas 90,000 kg of forage—just to get 5 km outside of Hatun Xuaxa toward the coast.
Environmental factors such as gorges or deserts constrain the flows that are possible from point to point. ...
The three models of centrality presented above provide a means to calculate the favorable network position of administrative centers, productive enclaves, and storage sites related to both production and administration. The results indicate that in the highlands the Inka built grand administrative centers and large storage facilities at regions of high and low centrality and large storage facilities without administrative centers at regions of low centrality, based upon the requirements of staple finance. The results also indicate that the Inka built smaller administrative centers with little or no storage capacity at regions of low centrality, based on the requirements of wealth finance. These findings are suggestive but preliminary. They await confirmation or disconfirmation based upon new archaeological and archival research that focuses on the communicative and exchange links between points in the road network. Such research would in turn allow for a more complete model of Inka exchange relations, which would incorporate secondary roads and centers and include elements of ocean travel. It would also provide the basis for a better understanding of the global patterns of the Inka political economy.
1 The story of Inka expansion has been told and retold a number of times. For early Spanish accounts see, e.g., Betanzos 1968 , 1987 ; Cobo 1964 ; Sarmiento de Gamboa 1965 ; Garcilaso de la Vega 1963–5 ; Cieza de León 1967 . For modern summaries, see Rowe 1946; Conrad and Demarest 1984; Rostworowski 1988; Patterson 1991; and Bauer 1992.
2 As Morris (1992) citing Burton (1966) points out, a metric ton of potatoes can be stored in a space as small as 1.6 cubic meters.
3 Perhaps the Inka northward expansion accounts for the storage complexes at the northern margins of the Inka empire. But the north’s relative independence from Cuzco should not be ignored as a contributing factor. Quito, for example, which was relatively independent from the influence of Cuzco, was an appropriate site for Atahuallpa to stage a war against his half brother, Huáscar, who had been crowned king in Cuzco just prior to the Spanish arrival.
5 Cobo’s (1964 : 126–7) six most important lateral roads, apparently in order of importance, are 1) Cuzco to Arequipa region, 2) Tumbes to sierra, 3) valley of Trujillo to Cajamarca, 4) valley of Paramonga to Xuaxa, 5) valley of Lima to Xuaxa, and 6) through Chuquiabo from the seacoast to the provinces of Chunchos.