A fast estimation procedure for discrete choice random coefficients demand model

We document speed-up gains of graphical processing unit (GPU) computing over central processing unit (CPU) for the estimation of discrete choice random coefficient demand model. When we use a moderate-sized GPU, the computation is six to twenty times faster, where the smallest speed-up factor, six, is obtained from a comparison with the parallel computing over sixteen CPU cores.     

A general-equilibrium intertemporal model of an open economy

Summary This paper develops a very general (general-equilibrium) intertemporal model of a country engaged in international trade which can be used to address a wide variety of issues of interest — in particular, econometric application — under the assumption that prices of tradable commodities (consumer goods and capital goods) and the interest rate are exogenous to the country. It allows for an arbitrarily large number of commodities which are distinguished into seven categories and for finite or infinite periods of time. This model can be used to draw various policy conclusions. We investigate how current net imports, the balance of payments on current account, current consumption expenditure, next-period bondholdings, current wealth, and current internal prices will react to exogenous changes in current external prices, the current interest rate, current taxes, current factor endowments, and current-period bondholdings. This paper also considers the integrability of net-import demand functions.Work supported by National Science Foundation grant SES-8607652. We wish to thank the editor and an anonymous referee for useful comments and suggestions.     

The choice of a foreign price measure in a Bayesian estimated new-Keynesian model for Israel

We estimate a small DSGE model by full information Bayesian techniques on the basis of Israeli data from 1995 to 2006. The model was first developed and estimated by means of classical GMM in Argov and Elkayam (2010), and since then it has been used at the Bank of Israel for monetary policy analysis. It is widely believed that in 2007 (out of sample year) as elsewhere worldwide, inflation rose in Israel due to high commodity prices in global markets. However, our baseline model attributes most of the high inflation in 2007 to supply shocks. One conjecture is that this model's result derives from the inappropriate original use of the unit value of imported consumer goods (which do not include unprocessed food and energy) as the main foreign price measure. We test this conjecture by re-estimating the model with various other foreign price measures that typically do reflect the global rise in commodity prices and compare the log-marginal likelihoods. We find that no other price measure outperforms the original choice in the sample period. Only the foreign trade-weighted CPI equals the performance of the original choice while improving the 2007 interpretation of inflation, and should therefore be considered the main foreign price measure. The proposed methodology for comparing the suitability of alternative measures for observable variables can be applied to any model with exogenous variables that are characterized by univariate equations.     

Optimal taxation of car ownership, car use and public transport: Insights derived from a discrete choice numerical optimization model

We formulated and numerically evaluated a model of car ownership, car use and public transport use for peak and off-peak hours of the day. The model was used to study the optimal tax structure for passenger transport in Belgium, with special emphasis on the optimal tax treatment of diesel versus gasoline cars. We obtained a number of interesting results. First, if the government can set all fixed and variable transport taxes optimally, the higher marginal external cost of diesel use implies that the optimal tax per kilometre for the use of a diesel car is higher than for the use of a gasoline car. Moreover, high congestion implies that the taxes on car use in the peak period are more than twice their current levels. However, the optimal tax on ownership of a diesel car is some 200€ below its current level. Second, if the government uses kilometre taxes that do not differentiate between fuel types, the optimal ownership tax on a diesel car is twice as high as the tax on a gasoline car. Furthermore, if political constraints restrict user taxes to their current levels, we find that optimal ownership taxes on diesel cars double, whereas those on gasoline cars rise by 30%. Finally, subsidies to public transport are found to be optimal as long as variable car taxes are not differentiated between periods.     

Impact of fiscal policy in an intertemporal CGE model for South Africa

This paper uses an intertemporal computable general equilibrium model to investigate the consequences of an expansive fiscal policy designed to accelerate economic growth in South Africa. A key contribution is made to existing literature on the transmission mechanism of fiscal policy in African economies. To the best of our knowledge, no published study has empirically analyzed the macroeconomic effects of fiscal policy in the context of an open, middle-income sub-Saharan African economy like South Africa using an integrated intertemporal model with such disaggregated production structure. The paper shows that an expansive fiscal policy would have a temporary impact on gross domestic product (GDP) but would translate into higher debt relative to GDP. Using increased taxation to finance the additional spending would lessen this impact but would also negatively affect macroeconomic variables. Increased investment spending would improve long-term GDP, under any financing scheme, and would decrease debt-to-GDP ratio as well as deficit-to-GDP ratio. This outcome is driven by the positive impact infrastructure has on total factor productivity. Sensitivity analysis shows that these conclusions are qualitatively similar for wide values of the elasticity of the total factor productivity to infrastructure. In fact, the conclusions hold even when comparing different financing schemes.     

Assessing the value of museums with a combined discrete choice/count data model

This article assesses the value of Dutch museums using information about destination choice as well as about the number of trips undertaken by an actor. Destination choice is analysed by means of a mixed logit model, and a count data model is used to explain trip generation. We use a utility-consistent framework in which the discrete choice model for destination choice is linked to an indirect utility function. The results are used to compute the compensating variation of particular museums and of the total group of museums in the sample.     

The demand for corporate jets: a discrete choice analysis

Discrete choice analysis is extended into the corporate arena with an analysis of general aviation business jet purchases. Based on previously overlooked data derivable from US aircraft registrations, an estimated model confirms that the business jet market processes information about product attributes similar to household markets for automobiles. Corporate purchasing decisions seem most sensitive to operating efficiencies, interior comfort, and brand loyalty. Price matters, but the average estimated response is inelastic. Other statistically significant attributes with choice elasticities less than one, include aircraft speed and reliability/safety.     

Testing for rationality: the case of discrete choice data

When working with micro data on consumer demand, there are many different situations where decisions involve only discrete choices. In this context, conditions under which an underlying rational preference structure exists are derived. Moreover, by introducing flexibility into the model, it is possible to identify nonrational behavior in the sample.     

A simultaneous discrete choice model of labor supply and wages for married women in Switzerland

A simultaneous model of female labor supply and wages is estimated. The labor supply model is formulated as a trichotomous discrete choice model in order to take into account the bimodal distribution of observed working hours. For women without observed wages the probability of choosing the observed work category is integrated over all values of the wage. The model is estimated using a sample of married women in Switzerland. The budget constraint is constructed by computing the disposable income for each hours point, taking into account the Swiss income tax system. The estimation results imply relatively modest wage elasticities of expected labor supply for working women. On the other hand the elasticity of the participation probability with respect to wages for nonworkers is twice as large. The magnitude of the income elasticities is comparable to other studies. The wage elasticities obtained with two alternative estimation procedures are considerably lower. These results show that wage elasticities crucially depend on how the problem of unobserved wages is solved.     

Bayesian estimation of duration models: An application of the multiperiod probit model

The multiperiod probit model is presented and Bayesian estimation using the Gibbs sampler with data augmentation is described. As an empirical illustration, the multiperiod probit model is then used to estimate a duration model using employment duration data for the Canadian province of New Brunswick. Bayesian estimation with unobserved heterogeneity is shown to be a simple extension of estimation of a duration model with no unobserved heterogeneity. More importantly, perhaps, some numerical problems encountered by other authors using Maximum Likelihood are avoided.I would like to thank M. Baker, G. Koop, A. Melino and D. Poirier for helpful comments and discussions. I would also like to thank M. Baker for providing his data.     

A note on an intertemporal cost function for a class of dynamic problems

Some properties of an intertemporal cost function for isoperimetric problems are analyzed. It is demonstrated that the cost function will be strictly decreasing in the length of the horizon T, and under certain conditions, strictly convex in T.     

An econometric model of tenure choice and demand for housing as a joint decision

This paper models a household's choice of tenure and demand for housing services as a joint decision imposing the restriction that both discrete and continuous decisions are derived from a single preference ordering. The utility index for households is the translog form of the reciprocal indirect utility function allowing for random preferences. Each household chooses between the two main tenures, the owner-occupied and subsidised rental sectors, but households may be rationed in either or both of these sectors and refused admission, in which case they are assumed to enter the third sector, uncontrolled rental. The model is estimated on UK data for 5895 households.     

A method for constructing an intertemporal utility function from a Friedmanian consumption function

It is shown that the intertemporal utility function can be reconstructed from data on demands if they are restricted to be linear in wealth, or Friedmanian.     

Expectations and intertemporal separability in an empirical model of consumption and investment under uncertainty

An intertemporal model of consumption and investment under uncertainty is formulated, and compared with the existing literature; it is argued that an assumption of myopia is necessary for its empirical applicability. It is estimated by maximum likelihood with quarterly British data. A specification search for a satisfactory form of expectations is made, and the estimated model is compared with a static demand system. Strong intertemporal separability is formulated as a nested hypothesis, and strongly rejected by a likelihood ratio test.     

Periodic and chaotic programs of optimal intertemporal allocation in an aggregative model with wealth effects

Summary We examine a discrete-time aggregative model of discounted dynamic optimization where the felicity function depends on both consumption and capital stock. The need for studying such models has been stressed in the theory of optimal growth and also in the economics of natural resources. We identify conditions under which the optimal program is monotone. In our framework, the optimal program can exhibit cyclic behavior for all discount factors close to one. We also present an example to show that our model can exhibit optimal behavior which is chaotic in both topological and ergodic senses.We would like to thank V. Bala, J. Benhabib, P. Dasgupta, R. Day, L. W. McKenzie, K. Nishimura, S. Roy, and a referee of this journal for comments on earlier versions of this paper. Research support from the National Science Foundation is gratefully acknowledged.     

Societal preferences for fertility treatment in Australia: a stated preference discrete choice experiment

Objective: To investigate preferences for fertility treatment from the Australian general population with the aims of calculating the willingness to pay in tax contribution for attributes (characteristics) that make up treatment and for an “ideal” fertility treatment program. We also assessed whether willingness-to-pay varies by the relationship status or sexual orientation of the patient.

Methods: A stated preference discrete choice experiment was administered to a panel of 801 individuals representative of the Australian general population. Seven attributes of fertility treatment under three broad categories were included: outcome, process, and cost. Attributes were identified through published literature, focus group discussions, expert knowledge, and a pilot study. A Bayesian fractional experimental design was used, and data analysis was performed using a generalized multinomial logit model. Further analyses included interaction terms and latent class modeling.

Results: Six of the seven attributes influenced the choice of a treatment program. Under process attributes, individuals preferred: continuity of care of clinic staff, where patients are seen by the same doctor but different nurses at each visit; “alternative” treatments being offered to all patients; and onsite clinic counseling and peer-support groups. Personalization and tailoring of the treatment journey were not important. Among outcome attributes, the improved success rate of having a baby per cycle and significant side-effects were considered important. Cost of treatment also influenced the choice of treatment program. Individual preferences for fertility treatment were not associated with patients’ relationship status or sexual orientation. Latent class modeling revealed sub-groups with distinct fertility treatment preferences.

Conclusion: This study provides important insights into the attributes that influence the preferences of fertility treatment in Australia. It also estimates socially-inclusive willingness-to-pay values in tax contributions for an “ideal” package of treatment. The results can inform economic evaluations of fertility treatment programs.     

A Bayesian evaluation of an efficiency-wage model with indeterminacy

ABSTRACT

This paper provides a quantitative assessment of a general equilibrium economy with non-Walrasian labour markets. Indeterminacy in the economy does not require production externalities or increasing returns but it rests on replacing the labour supply curve by a no-shirking condition on the efficiency-wage labour markets. The model is estimated on U.S. data via full information Bayesian methods. The shirking model is capable of matching several stylized facts of the aggregate economy and the labour market. Data favour a version of the artificial economy that is characterized by determinacy.     

Spectral analysis as a tool for model discrimination: An application to international capital movements

Empirical discrimination between ‘stock’ and ‘flow’ formulations of international investment based on regression tests has consistently proved unsuccessful. By applying the new technique of model discrimination in the frequency domain, this note provides unequivocal empirical support for the stock model.     

Bayesian estimation of the output gap for a small open economy: The case of Canada

We introduce a novel approach for estimating output gaps for small open economies. Identification is based on a multivariate trend-cycle decomposition in which transitory exchange rate movements are linked to the output gap and inflation. The model is then applied to Canadian data.