Econometric Prediction Research Articles

Targeting humanitarian aid using administrative data: Model design and validation

We develop and assess the performance of an econometric prediction model that relies on administrative data held by international agencies to target over $380 million annually in unconditional cash transfers to Syrian refugees in Lebanon. Standard metrics of prediction accuracy suggest targeting using administrative data is comparable to a short-form Proxy Means Test, which requires a survey of the entire target population. We show that small differences in accuracy across approaches are largely attributable to a few data fields. These results are robust to a blind validation test performed on a random sample collected after the model derivation, as well as the type of estimator used for prediction. We discuss relative costs, which are likely to feature prominently when alternative approaches are considered in practice.

Improving econometric prediction by machine learning

ABSTRACT We present a Machine Learning (ML) toolbox to predict targeted econometric outcomes improving prediction in two directions: (i) by cross–validated optimal tuning, (ii) by comparing/combining results from different learners (meta–learning). In predicting woman wage class based on her characteristics, we show that all our ML methods’ predictions highly outperform standard multinomial logit ones, both in terms of mean accuracy and its standard deviation. In particular, we set out that a regularized multinomial regression obtains an average prediction accuracy almost 60% larger than that of an unregularized one. Finally, as different learners may behave differently, we show that combining them into one ensemble learner proves to preserve good predictive accuracy lowering the variance more than stand-alone approaches.

Now-And Forecasting Unemployment from Google's Job-Search Activity

Big data trends, have gained popularity among practitioners for its potential applications on accurate econometric prediction. This paper presents an application where unemployment job search indicators for the G7 countries are based on Google trends data. For each country, a set of google search keywords is identified, then a data reduction technique is applied to estimate our unemployment indicator which is considered a latent common trend. Finally the predictive performance of our indicators is assessed by an automatic stochastic specification with respect to unemployment levels and other key labour market outcomes.

Distance Decay in the Willingness to Pay for Wine: Disentangling Local and Organic Attributes

This paper investigates how the residents of a French wine-producing region value the attributes of wine. We elicit the willingness-to-pay (WTP) for organic/non-organic and local/non-local wines with increasing levels of information on the impact of agricultural practices at both global and local scales. The analysis shows that there is a significant organic premium associated with both local and non-local wines. This organic premium significantly increases with information and significantly decreases with the distance between the consumer’s home and the vineyard. Based on the econometric predictions of the WTP, we show that a per-unit tax on non-organic wines or a standard imposing organic practices increases welfare through the internalization of the attributes revealed by the experiment.

El Paso Housing Sector Econometric Forecast Accuracy

There is comparatively little empirical evidence regarding the accuracy of regional housing sector forecasts. Much of the recent analysis conducted for this topic is developed for housing starts and indicates a relatively poor track record. This study examines residential real estate forecasts previously published for El Paso, TX using a structural econometric model. Model coverage is much broader than just starts. Similar to earlier studies, the previously published econometric predictions frequently do not fare very well against the selected random walk benchmarks utilized for the various series under consideration.

El Paso Housing Sector Econometric Forecast Accuracy

There is comparatively little empirical evidence regarding the accuracy of regional housing sector forecasts. Much of the recent analysis conducted for this topic is developed for housing starts and indicates a relatively poor track record. This study examines residential real estate forecasts previously published for El Paso, TX using a structural econometric model. Model coverage is much broader than just starts. Similar to earlier studies, the previously published econometric predictions frequently do not fare very well against the selected random walk benchmarks utilized for the various series under consideration.

Regional econometric income forecast accuracy

AbstractEconometric prediction accuracy for personal income forecasts is examined for a region of the United States. Previously published regional structural equation model (RSEM) forecasts exist ex ante for the state of New Mexico and its three largest metropolitan statistical areas: Albuquerque, Las Cruces and Santa Fe. Quarterly data between 1983 and 2000 are utilized at the state level. For Albuquerque, annual data from 1983 through 1999 are used. For Las Cruces and Santa Fe, annual data from 1990 through 1999 are employed. Univariate time series, vector autoregressions and random walks are used as the comparison criteria against structural equation simulations. Results indicate that ex ante RSEM forecasts achieved higher accuracy than those simulations associated with univariate ARIMA and random walk benchmarks for the state of New Mexico. The track records of the structural econometric models for Albuquerque, Las Cruces and Santa Fe are less impressive. In some cases, VAR benchmarks prove more reliable than RSEM income forecasts. In other cases, the RSEM forecasts are less accurate than random walk alternatives. Copyright © 2005 John Wiley & Sons, Ltd.

Forecasting residential burglary

Following the work of Dhiri et al. [Modelling and predicting property crime trends. Home Office Research Study 198 (1999). London: HMSO] at the Home Office predicting recorded burglary and theft for England and Wales to the year 2001, econometric and time series models were constructed for predicting recorded residential burglary to the same date. A comparison between the Home Office econometric predictions and the less alarming econometric predictions made in this paper identified the differences as stemming from the particular set of variables used in the models. However, the Home Office and one of our econometric models adopted an error correction form which appeared to be the main reason why these models predicted increases in burglary. To identify the role of error correction in these models, time series models were built for the purpose of comparison, all of which predicted substantially lower numbers of residential burglaries. The years 1998–2001 appeared to offer an opportunity to test the utility of error correction models in the analysis of criminal behaviour. Subsequent to the forecasting exercise carried out in 1999, recorded outcomes have materialised, which point to the superiority of time series models compared to error correction models for the short-run forecasting of property crime. This result calls into question the concept of a long-run equilibrium relationship for crime.

The Response of Household Consumption to Income Tax Refunds

A central implication of the life-cycle (or permanent-income) theory is that consumption should not respond to predictable fluctuations in income. Tests of this implication have yielded mixed results, especially on micro data (Angus Deaton, 1992; Martin Browning and Annamaria Lusardi, 1996). In large part this might be due to the difficulties of isolating the predictable component of income at the micro level. Most tests proceed by instrumenting for income, but since the available instruments are typically poor, such tests might be prejudiced against finding significant excess sensitivity of consumption to income (John Shea, 1995).' Also, it is not clear how closely the resulting econometric predictions of income coincide with agents' actual expectations of income. To avoid these difficulties this paper examines the response of household consumption to a particular type of income that is both predictable and transitory-income tax refunds. Since a refund depends on events in the previous calendar year, it is predictable income as regards consumption in the year of its receipt. Consequently, under the life-cycle theory consumption should not increase on receipt of a refund.2 In addition to testing the canonical model of consumption, this paper provides estimates interpretable as the marginal propensity to consume (MPC) out of refunds. Since federal tax refunds now amount to over $80 billion per year (averaging well over $1,000 per refund), these estimates are of interest in themselves. More generally they bear on the impact of even preannounced and temporary changes in fiscal policy. The paper begins by surveying related studies in Section I. Section II describes the data, the Consumer Expenditure Survey (CEX), which of the leading U.S. micro data sets has the most comprehensive coverage of expenditure. The empirical specification is set out in Section HI. Section IV reports the results, and Section V concludes.

Government Policy, Saving and Investment

The author examines the role of government in its influence upon saving and investment and disputes the argument that past government policy with respect to income redistribution, high marginal tax rates, and social security have depressed national saving. With respect to investment he argues that government policy over the past thirty years has encouraged private investment and that currently more attention should be given to public investment in research, roads, health, and all forms of human capital, including job training and education generally. Although many of the arguments in the article are subtle, reference to a prior specific econometric prediction of the effects of tax cuts upon investment permits the author to say “I told you so.”

Assessing the Quality of Prediction Baykal: Bayesian Prediction by Kalman Filtering

In Bayesian statistics for prediction a complete predictive density is calculated. This is an ideal means to asses the quality of prediction, as arbitrary measures of dispersion can be computed. The predictive density not only reflects the randomness of the residuals, but also takes into account the uncertainty of parameter estimates. In this paper a method (BAYKAL) is presented, which computes the predictive density by a Monte Carlo integration technique using the Kalman filter. BAYKAL is applicable to linear models with various features including simultaneous equations, observation errors, and correlated residuals. Apriori informations can be incorporated in a flexible way. Complete marginal densities, various parameters of these densities and measures of numerical errors can be obtained. For three small models the BAYKAL - predictions are compared with usual econometric predictions. The results indicate that the dispersion of BAYKAL predictions are always larger than the dispersion of the corresponding econometric predictions, which neglect the uncertainty of the parameters.

Daily Fluctuations in Campground Use: An Econometric Analysis

AbstractMany recreation activities are characterized by fluctuating use levels over time. This paper presents a theoretical and quantitative analysis of fluctuations in campground use series. Theoretical use functions are specified as the reduced form of a demand and supply system, where the derived demand for campground use is based upon the household production approach to consumer demand theory. A pooled time‐series cross‐section econometric use model is specified, and its estimated parameters are presented. In addition, campground use series are modeled incorporating Box‐Jenkins time‐series analysis techniques. Econometric predictions are more accurate relative to autoregressive‐integrated‐moving‐average (ARIMA) predictions.