Learning Outcome:  After this lesson, you will be able to evaluate whether a SAM needs to be updated.

Why Do Some SAMs Have Older Data?

Social Accounting Matrices (SAMs) describe the economic activity in an economy at a point in time, usually a year. The US333 SAM describes the US economy in 2007. SAMs typically date to a period of time that may be several years in the past.

Why do SAMs have relatively old data?  There are two main reasons.  One is that a SAM requires a comprehensive set of data, including macro data, trade data, tax policy data, and microeconomic data on the commodity composition of demand.  All components of the SAM data set must be assembled to construct the SAM.  Because some data may only be published after a delay, the construction of the SAM is constrained to the latest year that all data become available.The second reason is that - once data are available - creating a SAM is a major research undertaking, especially if it is part of a balanced global database with many countries, such as the global GTAP database. 

Structural Changes

Older data may not be a problem because CGE models are "structural."  This mean that it is not the level of activity that matters; it is structure that matters. Structure refers to the shares of economic activity. For example, the share of output that is exported, the share of consumption that is imported, and the share of workers employed in each industry, are all structural characteristics. We explore the structure of the US economy in our lesson on structure tables in this module.  

Is it always OK to have a relatively old SAM?  That depends. If the economy's structure has not changed, it is OK.  If the structure has changed, the SAM should be updated.  For example, perhaps petroleum accounted for 50 percent of a country's total exports in 2015.  If it continues to account for 50 percent of exports in 2022, the SAM may not need an update. But, if the petroleum supply has been exhausted and services have become the main export, the SAM should be updated to reflect this new economic structure.   

The approach to updating the structure of the economy is to introduce the new data as a model experiment and allow everything else in the economy to adjust.  In our example, the new trade data is introduced as an experiment, and production, employment and consumption patterns are allowed to adjust accordingly.  Perhaps petroleum workers move into the services sector, and consumers purchase less petroleum. The SAM that results from the experiment will have a new structure, and becomes the new, updated SAM. Before using it in the model, the modeler needs to check that the new SAM provides a good representation of the current structure of the economy.    

Better Policy Information

What if the modeler acquires better information on policies?  As an example, let's assume that their SAM describes import tariffs with a uniform rate of 5 percent. Then, the government publishes detailed information on tariffs, which actually vary a lot across commodities.  If the modeler is analyzing a scenario such as trade liberalization using the uniform tariffs, the results will be different than if the modeler accurately captures the varying import tariffs that are actually imposed. When better information on policies becomes available, the SAM should be updated.

The approach to updating policies in the SAM database is to introduce the new policy as a model experiment but hold everything else in the economy constant to the extent possible.  One way to do this is to introduce the new policy data as an experiment while adjusting elasticities in a way that reduces any impact on the original structure. For example, the GTAP model has a utility that helps researchers do this update by redefining elasticities so that the utility and production functions become Cobb-Douglas, in which budget shares and factor shares remain constant. The SAM that results from this policy experiment becomes the new, updated SAM to be used in the model.  

 _{Copyright: Cornerstone CGE CC 4.0  BY-NC-SA}