Estimation of consumer demand systems with binding non-negativity constraints

Are Underground Workers More Likely to be Underground Consumers?

“…24 However, given that the EF estimator uses more information than the CEF estimator since the former uses the full sample while the latter uses a truncated sample , one should expect, on intuitive grounds, the EF estimator to be asymptotically more ecient than the CEF estimator. However, Stapleton and Young 1984 were unable to show that this is necessarily the case.…”

“…12 Barter of goods and services produced informally is not modeled since it represents less than 6 of the size of underground economy in our sample. 13 On the contrary, it is clear, from Kuhn-Tucker conditions, that the functional form for the unconditional demand for irregular commodities will vary depending upon whether the individual works or not in each market, which makes the estimation of the model much more complex to perform e.g., see Wales and Woodland 1983 . 8 max fCr;Cug U C r ; C u ; h r ; h u ; z 3 ; 6 subject to P r C r + P u + c h u ; z 2 C u = m; where m is the level of total expenditures on commodities, as given by the right-hand side of equation 4 , 14 and subject to the non-negativity constraints on C r and C u . The rst-order conditions for an interior solution to this program yield the conditional demand for undeclared goods and services: 15 C u = C u P r ; P u + c h u ; z 2 ; m; h u ; h r ; z 3 :…”

Multi-Category Competition and Market Power: A Model of Supermarket Pricing

“…The UK's Expenditure and Food Survey (EFS) includes a variable for gross current household income (variable p352). We estimate household income by regressing this income variable (for years [2003][2004][2005]) on other demographic variables in the ESF that map to those in the TNS survey, namely indicator variables for the number of cars (0, 1, 2, ≥ 3), adults (1, 2, ≥ 3) children (0, 1, 2, ≥ 3), household size (1, 2, ..., ≥ 6), geographic region in Great Britain (10 regions), social class (6 classes as described in Appendix C), tenure of residence (dummies for whether the home is privately owned, privately rented, or public housing, structure of residence (detached house, semi-detached/terrace, and apartment), year, sex of the Household Reference Person (HRP), and age of the HRP (≤24, [25][26][27][28][29][30][31][32][33][34][35][36][37][38][39][40][41][42][43][44][45][46][47][48][49][50][51][52][53][54],55-64,≥ 65) We dropped the top and bottom 1% household incomes to avoid outliers. The R 2 is 0.51 and the number of observations in the regression is 17, 335. yielding 180,000 observations.…”

“…(We do the individual weighting at the product group g level but not the individual product h level because many individual products such as "private labels" are firm-specific and an individual consumer typically only visits a subset of the firms in the data). (53) where weights ω i g are now the total expenditure share (over the three year period) of each product group g by consumer i and satisfy g∈G k ω i g = 1 for each i. The weights are constant across stores and over time.…”

A joint model for the perfect and imperfect substitute goods case: Application to activity time-use decisions

“…The non-linear utility structure used in Bhat's approach was employed originally by Kim et al (2002) as a specific satiation-based formulation within the broader Kuhn-Tucker multiple-discrete economic model of consumer demand proposed by Wales and Woodland (1983). 1 Bhat's model, labeled the multiple discrete-continuous extreme value (MDCEV) model, is analytically tractable in the probability expressions and is very practical even for situations with a large number of discrete consumption alternatives, unlike the models of Wales and Woodland and Kim et al In fact, the MDCEV model represents the multinomial logit (MNL) form-equivalent for multiple discrete-continuous choice analysis and collapses exactly to the MNL in the case that each (and every) decision-maker chooses only one alternative.…”