Obtaining unit values

[Moses Muwonge]

Hi statalist users,

My call goes out on how to obtain the unit values for each food group prior to estimation of the linear approximation to almost ideal demand system. With the data set am working with, I have got variables reporting on both the quantity & values for the (consumption out of purchase, consumption out of home produce, and that received in kind/ free).
Okay, I have categorized the individual food items in to 13 food groups. Could someone be knowing the best way unit values for each of the 13 food groups can be generated. Am faced with a situation where i need to regress each of the unit values for the 13 food groups with the selected independent variables in order to predict the corresponding prices for each of food group but so far the approach i have employed also predicts negative prices for which if am to transform the predicted prices into log prices give me an error which will obvious creates inconsistencies in parameter estimation.

Thank you

[Phil Bromiley]

You didn't get a quick answer. You'll increase your chances of a useful answer by following the FAQ on asking questions. Also, note that we are not all from your area so I and many others won't know what "linear approximation to almost ideal demand system" means.

Your post is extremely hard to understand. I have no idea what the structure of your data looks like, and I don't know what you mean by unit values. When you say "obtain" I don't know exactly what this means either - find new data, estimate predicted value, or what. Explicit equations might help as well.

[Moses Muwonge]

Hello Phil
the structure is as below

Out of purchase				Home production		In kind received
From home		Away from home
Ceb06	Ceb07	Ceb08	Ceb09	Ceb10	Ceb11	Ceb12	Ceb13	Ceb14	Ceb15
1	2800		100		100			2800
1.2	12000				120			6000
.14	6000							3000
						12.5	6000	12000
.21	3500							500
				2.1	21000				500

where ceb06, ceb08, ceb10 & ceb12 are columns for quantity. ceb07, ceb09, ceb11 & ceb13 are columns for value, then ceb14 is the market price, ceb15 a column for the farm gate price.
That is Unit value = Expenditure/quantity for instance of a given food item.
Am analyzing Micro data using a cross sectional data set for the household survey with rows identifying household members. There additional columns not display in the structure reporting household characteristics like age, gender education, the other represent the individual food items listed by the household in that the household ID is duplicated among a list of households members represented row wise.

Remember a household might have consumed out of purchase or out of home produce or received that food item in kind or this could be a combination of the two or more. For this case am to estimate a demand system called "linear approximation to almost ideal demand system" (LA/AIDS).So am tasked with generating a column to represent the unit value(unit price) for each food item consumed by that household. since i can not base on the addition of the market price & farm gate price price, the two are different. I need unit price column in which case am looking for the best way to represent prices in that the survey did not provide.

Then i have a problem on how to construct the corresponding prices after grouping the food individual food items . For example ( rice, wheat, corn, maize) in to a food group i have named cereals
I need to use the dependent variable for estimation of the price equation of the form
p_i = per capita food expenditure + household size + years spent schooling + gender + residence dummy(urban) + region dummies

In this case i will have to regress unit price for the group on to the household social-economic, demographic characteristics . This will help me generate prices for those household who never reported expenditure on some food item .and to avoid the zero reported value prior to the estimation of parameters for (LA/AIDS) at the next stage grouping of the food items.After categorizing the individual food items e.g ( into 13 categories ). Each of the predicted food price column will represent a particular food category.

Thanks