Look in the HRM journals. That is why SEM (Structural Equation Models) are taught in business schools. SEM (simultaneous equation models) are the econometric topics. See the econ journals. Both are taught currently in our Graduate School of Management. If I were to wonder about your question I would do a literature review not an armchair survey. D. Booth

Structural Equation Modeling is used so popularly in researching human behaviours. if you want to know the importance and application of SEM, you can follow some scholarly journals such as: Econometrics, Biological and Ecological Research, Business or Economics.

As David Eugene Booth said, structural equation modeling is not taught in econometrics. This is the foundation of immense criticisms of some scholars

Chen, B., & Pearl, J. (2013). Regression and causation: A critical examination of six econometrics textbooks. Real-World Economics Review, 65, 2-20.

Best,

Holger

Many thanks to David Eugene Booth , Huynh Tan Nguyen , and Holger Steinmetz for the useful advice.

This is will sound rude, sorry!

I am always amazed how intellectually isolated much of economics is. I do not use much SEM myself but I see it all the time in journals and presentations throughout the social sciences. A good sign of the scale of this the development of specialist software AMOS, MPlus and the creatively recently introduced SEM capability in Stata which economists do use. And of course, as Holger Steinmetz says, there is all the Pearl stuff on causal inference, which should be of huge interest to economists.

First of all, many thanks to those who shared their views in this discussion.

I think there is a little confusion between structural equation modeling and simultaneous equation models in the discussion. I do agree that structural equation modeling is widely used approach in various disciplines with wide-range of applications.

The question here is on practical applications on simultaneous equation models. These models also use the term structural, or behavioral, equations because they may portray the structure of an economy. Of course, simultaneous equation models establish theoretical relationships among variables. A concern is how these models are estimated and analyzed as shown in conventional Econometrics textbooks.

For example, if one knows a particular theoretical model (e.g., Keynesian model of income determination) is under simultaneous equation models then, according to the books, in the first step, one needs to check for identification whether a model is identified or unidentified using the order or rank condition. Usually, these two conditions suggest contradicting results because one is sufficient condition another is necessary and sufficient condition. Based on the identification result, one chooses the proper technique to estimate simultaneous equation models.

All in all, I think my question should be rephrased as follows:

Could anyone suggest any recently published article, material, or report that uses simultaneous equation model that is analyzed based on identification properties?

Ok that is clear.

Here is one of interest

Steele, Fiona, Vignoles, Anna and Jenkins, Andrew (2007) The effect of school resources on pupil attainment: a multilevel simultaneous equation modelling approach. Journal of the Royal Statistical Society, series A, 170 (3), pp. 801-824

If you do not do the simultaneous bit, you find that extra resources lead to worse attainment as it is the poorest achieving schools that get the most money. What is methodologically important here is that attainment is at the level of the pupil, and the resources and the instruments at the level of the school and or district.

This simultaneous model can be specified and estimated as a multilevel multivariate variate response model with more than one response variable (they can be discrete or continuous) and can involve further levels: see

http://www.bristol.ac.uk/cmm/software/mlwin/mlwin-resources.html#multipleresp

Thank you Kelvyn Jones for sharing the paper. The paper looks interesting !!