Economic multipliers for Colombia: Spatial distribution of the impacts (Do we have a system of cities?)

[+] Juan C. Duque1 , Alejandro Betancourt2 , Christian Posso3

Welcome to RiSE's new interactive tool for visualizing the interindustry dynamics in Colombian economy. This tool has been developed in collaboration with UN-Habitat with the aim to contribute new elements to the initiative “National Urban Policy within a National Spatial Framework.” RiSE team thanks Doctor Elkin Velasquez (Chief, Regional and Metropolitan Planning Unit. UN-Habitat) for his insightful comments and support during the implementation of this tool.

Data source

We use the product utilization matrix for Colombia 2006 (base 2005) prepared by the National Administrative Department of Statistic (DANE) to measure the impact on the global economy that results from a demand shock of 1.000 million Colombian pesos (COP), from 2011, in a given industry (This impact is known as Leontief Multiplier). The product utilization matrix uses a 61-industry classification based on the two-digit CIIU (rev.3) categories. The information about total employment is obtained from the Continuous Household Survey (first semester of 2006).


Economic impact

The graph below represents Colombian interindustry relationships, where each node symbolizes an economic industry. The links between nodes represent the most important interindustry relationships according to the algorithm for industry clusters identification devised by Duque and Rey (2008) (see also Duque, Rey and Gomez (2009) for more details about this methodology). The color of the nodes represents the value of the multiplier of each industry, which ranges from black to light red, such that the lighter the red the higher the value of the multiplier.


NOTE 1: Industry "Domestic Services" behaves as an "island" because it shows no relationships with other industries. Industry "Waste products" was removed because it reports no outputs, which results in a singular matrix. Market and non-market Education Services were merged, as well as market and non-market Social and Health services.


NOTE 2: All monetary values are reported as prices from 2011.


Spatial distribution of the economic impact

Once the economic impact is calculated we use information of employment by industry on each of the 13 metropolitan areas in Colombia to approximate how the impact is going to distribute across the system of cities. This relies on a number of strong assumptions like equal labor productivity and proportional distribution of the impacts (based on employment shares) across the cities; but it give us a rough idea of the short term impacts of a demand shock in the system.


The impact is then visualized in two maps. The map on the left shows the distribution of the direct impact; i.e., the distribution of the increment in the outputs of the impacted industry. The map on the right shows the distribution of the indirect impacts; i.e., The sum of the increment in the outputs of the industries that have commercial transactions with the impacted industry. Thus, the sum of the direct and indirect impacts across the system of cities equals the Leontief multiplier.


The way in which direct and indirect effects are distributed in the system can tell us some information about whether or not the cities are working as a system of whether they are acting as isolated cities. Some interesting patterns are:


A system of cities: In this case the direct impact tends to concentrate in a few number of cities (or even one city, in the case of high level of spatial concentration of a given economic activity); and the indirect effects is allocated in the same city or other satellite cities that act as suppliers. These types of models become feasible when the unit costs savings resulting from specialization exceed the unit transports costs (E. J., Gauthier H. L. and M. E. O'Kelly (1996) Geography of Transportation).


Unconnected cities: In this case the direct and indirect impacts are distributed across the system; i.e. each city satisfy their own needs; thus, connection within cities is not required and there are not economies of agglomeration (either, because inefficiencies in the system of because the product, or service, need to be located near to the consumer.


Unconnected industries: In this case the impacted industry has no economic transaction with other industries in the system. In those cases, the map of indirect impacts will be empty. Impacts on the “Domestic Services” industry shows this type of patterns.


To use this tool go through the following steps:


1. Find an industry (this step is optional):

You can locate an industry in the graph by choosing it from the following list:



The selected industry will stand out in the network. You can also identify an industry by locating the cursor on any node in the network. You will see an emerging box with the name of the industry.


2. See how a demand impact spreads in the global economy :

Doble-click on an industry (node) to simulate an increase of 1 thousand million in current COP from 2006 in the demand for that industry (hereafter COPs will be reported in current prices from 2006). Once an industry is clicked you will see how the impact spreads throughout the other industries in the economy.


Below the graph you will find more useful information about the industry that you are impacting.


Direct impacts by city

Indirect impacts by city

+ Publicaction and Consulting

Releated Publications

Take a look to some of the publications coauthored by our members in which we develop new methods for industry clusters identification.

Releated consulting projects

Consulting projects in which we combine techniques such as input-output analysis and system dynamics.

See all +

+ Information

Developer Team

Juan C. Duque, Alejandro Betancourt, Christian Posso.

Contact Us

Your comments and suggestions are welcome.

See all +