-quaids- warning

[Tony Johnson]

Hi,

I have two questions regarding the -quaids- command and, given that I am new to demand system estimation, any help would be greatly appreciated.

My version of Stata is Stata 13-MP on Windows 7. I am trying to estimate a demand system for 12 broad categories of goods (i.e. for the whole economy), using 6 years of Household Budget Survey data (2008-2013). In total, I have about 22,000 observations/households. CPI data is used as prices and all household from a particular year face identical prices.

My questions are the following:
1. Every time I run -quaids-, I get "Warning: variance matrix is nonsymmetric or highly singular" and thus no standard errors for the parameters are estimated. My understanding is that the warning is caused by the low variation in prices. Is this true?
2. My price data is in index form with the price of all categories normalised to 100 for the base year (2009). What form should prices be in?

Please let me know, if any of the above is not clear.

Thank you in advance for your help,
Tony

[May Abaza]

Dear Stata Users and dear Tony,
I am glad that I saw your post because I am having the same problem now.

I have Egyptian Household Income, Expenditure and Consumption Survey data. and I am applying the Quadratic Almost Ideal Demand System

The data has seven expenditure categories food,housing, health, education, transportation, cloth, and other as the seventh category
I have the per-capita total expenditure and I computed the expenditure share for each category.
The prices of each category are the same for all households.

When applying Poi's (2012) command QUAIDS ,I get the same "Warning:variance matrix is nonsymmetric or highly singular" message and missing standard errors in the output.

And both e -returned e(V) ( the variance-covariance matrix of the estimators), and e(Vest) (the variance-covariance matrix of estimated parameters)
are zero's.
So It is clear that not having varied prices by households is a problem, but they are what I have.
Does this mean that QUAIDS is not applicable is this case?


I couldn't find a reply to this post so I was wondering if you were able to figure it out.



I appreciate if there is an update for solving this issue.

Thank you
May