Javier Barbero Jiménez (p), Francesco Di Comite

Different methodological approaches have been proposed recently to introduce firm heterogeneity á la Melitz in computable general equilibrium models, thus identifying new channels of adjustment in the economy and providing additional depth to models based on a combination of Armington assumptions and Krugman's new trade theory. However, these extensions have been mainly tested with simulated data or applied to macroeconomic settings characterised by a scarcity of granular data and heavy reliance on distributional assumptions to feed the models. In this paper, we build upon the European Commission efforts to build up Europe-wide Social Accounting Matrices for every EU NUTS2 region and macrosector and show how a spatial computable general equilibrium model can be enhanced with a heterogeneous-firms module at different levels of spatial aggregation, discussing its data requirements and the computational challenges involved.

The explicit introduction of heterogeneous firms in the picture affects significantly the outcomes of simulations and highlights how aggregate macroeconomic figures can hide diverging outcomes at the firm level, thus improving the model relevance for policy design. The most notable example concerns trade flow adjustments after a policy shock, which is in this setting the result of individual decisions of firms across sectors and along the productivity distributions in ways that can actually partially offset each other. This means that even an increase in bilateral trade flows may be the result of the combination of heterogeneous outcomes, with some different firms expanding, shrinking or exiting the market as a result of the same macroeconomic shock. The dynamics generated by such a model are therefore closer to what can be observed empirically in the international trade literature at the firm level, lead to within- and between-sectors firms reallocations. With regard to the within-sector firm reallocation, policy shocks can lead the less productive firms to exit the market resulting in higher market-sector shares for highly productive firms, which in turn bid up input prices and affect firms in all the other sectors and territories. This feature can only be observed in a heterogeneous firms framework.

We present this point formally by showing the economic outcomes associated with stylised policy shocks under three different sets of model settings: traditional perfect competition with Armington differentiation; homogeneous firms à la Dixit-Stiglitz-Krugman; heterogeneous firms à la Melitz.

Sun-Young Heo , Tae-Heon Moon , Ju-young Kim (p)

The crime had become more complex and diverse than the past. In order to cope with this phenomenon, various new methods for preventing and reducing crime had been developed and been applied to the field. CCTV is one of powerful devices that had actively installed in urban space. However, with the expectation of CCTV, there is also a growing concern that CCTV will have little effect on crime prevention.
CCTV can result in ‘crime displacement’ in which criminals simply move in response to preventive activities. On the other hand, CCTV can have a positive effect of preventing crime even in neighboring areas (‘diffusion of benefits’). Thus, CCTV strategy needs to investigate first the effectiveness of CCTV. In order to analyze the positive and negative effects of CCTV on crime prevention, this study conducted an empirical analysis in S city, Korea.
For the analysis, the crime incidents in study area (1,070 cases in 2012, 1,112 cases in 2013, 1,240 cases in 2014) and CCTV (172 cases) installed in 2013 are converted into spatial data using QGIS. In study area, six regions were selected where the crime rate was high and the number of crime incidents before and after CCTV installation was comparable. Then, using the Weighted Displacement Quotient (WDQ) developed by Bowers & Johnson (2003), this study analyzed the effectiveness of CCTV quantitatively.
For the effectiveness measurement, study areas were divided into three sub-areas: experimental area, buffer area (transition area), and control area. The experimental area is the CCTV installation area and the buffer area is the transition area where the crime displacement is expected due to the CCTV installation in the experimental area while surrounding the experimental area. The control area is set as an area where the change of crime in the experimental area and transition area may not affect.
As a result, it was confirmed that theft incident in residential area was reduced. However, there was no or minimal effect of crime displacement in the surrounding area. Although the effects of CCTV in the surrounding area are not as much as the direct effects of crime reduction, it found that there is a possibility that crime could decrease.
The results of this study show that if CCTV installed in a place that does not have the effect of crime displacement and can spread the effect of crime control, it could prevent or reduce the crime.

Sung-goan Choi (p), Hana Kwon

The regional input-output model (RIO model) has been actively used in the formulation, evaluation, and forecasting of regional policies. In spite of its usefulness, the RIO model is not a simple way to estimate the regional input coefficients, and much time and cost is needed to create a survey-based RIO model. For these reasons, many previous studies have been researching non-survey techniques. When the conventional LQ methods are used, they tend to overestimate the benchmark table, as Round(1978), Zhao and Choi(2015) and Kwon and Choi(2017) pointed out.
Recently, the FLQ method proposed by Flegg et al. (1995), Flegg and Webber (1997) has been cited as an alternative in RIO modeling. However, the FLQ formulae are limited in identifying the leading industries using the backward linkages, because the rank correlations between the benchmark table and estimated RIOT are not high. Since the key sectors have high backward linkages in the whole economy, identifying the leading industries based on the objective evidence is important for local policymakers. In addition, recently there has been ‘jobless growth’ in some industries because of changes in productivity and technological progress. When ‘jobless growth’ takes place, the use of the employment data may distort the LQ, thus may overestimate or underestimate the input coefficients, and therefore, poses a problem about data use for accurate RIO modeling. Kwon and Choi (2017) proposed a data hybrid method (DHM) for regionalization technique. DHM is a way to significantly improve the problems of existing methods by applying the concept of employment elasticity to the RIO model. The attributes of the elasticity concept used require industry-specific employment and output information In order to construct a DHM, basic data, such as employment by industry, and output by sectors for two distinct periods of time, are required.
On the other hand, the weighted location quotient (WLQ) method proposed in this study uses basic data on employment or output (or value-added) by industry for a single year. By presenting the WLQ method as a more reliable and accurate non-survey method for the RIOT that can be applied universally in South Korea's regional analysis with constraints on available statistical data, this method can be used to analyze the output multiplier and to select government investment projects in the future. In this research, we proposed the WLQ as a more accurate method to estimate the RIO model and compared the empirical results with the previous methods

John Deely (p), Stephen Hynes , John Curtis

Random utility theory is founded on the concept that an individual selects the alternative that gives them the highest level of utility, given the individual’s preferences and perception of a good. Discrete choice analysis, however, seldom uses an individual’s perception of a good, instead, more convenient objective data are employed. This paper aims to explore the viability of objective data as a suitable replacement for subjective data in recreational site choice modelling. Random parameter logits are applied to coarse angling site choice data where two site attribute data sets are used; the first is comprised of users’ perception of the site attributes and the second is composed of fishery managers’ perspective of those same attributes. The results reveal that models based on the subjective data outperform those of the objective data. The derived welfare estimates indicate a divergence between the two sources of data in terms of the magnitude of the estimates but not direction. Further analysis is conducted to determine if the manager’s objective ratings are measuring the sites using a similar set of criteria as the user’s subjective ratings. The results suggest that the managers’ perspective is closely aligned with the anglers’ who frequent the sites most often.

Esteban Fernandez-Vazquez (p), Monica Serrano , Alberto Diaz Dapena (p)

Household Surveys (HS) report consumption figures that can be used to calculate consumption impacts when used in combination with IO models, but the samples on which HS base are rarely representative to produce reliable estimates at a detailed (sub-regional) spatial scale. This implies that the use of IO analysis when quantifying the impacts derived of household consumption typically limits to estimate them at an aggregated regional or national scale. In this paper we apply an econometric procedure to estimate consumption figures at a highly disaggregated geographical scale, which would allow for calculating impacts of changes in the household consumption in particular sub-regional spatial units (i.e., cities, metropolitan areas or municipalities). First, we base in the methodology developed in Elbers et al. (2003) and Tazzoni and Deaton (2009), which predict spatially disaggregated economic indicators by combining data from HS’s with the information contained in the Population Census (PC), since these datasets usually contain (non-economic) indicators observable at a highly detailed geographical classification. Then, a Generalized Maximum Entropy (GME) estimator is applied to adjust these initial estimates on consumption, making them consistent with the consumption aggregates reported on an IO table at industry level.

As an illustration, we study the micro-data of the Spanish Household Budget Survey and the Spanish Census of Population elaborated by the National Institute of Statistics (INE) in 2011. From these two databases we estimate consumption figures for more than 500 spatial units, which are subsequently adjusted to the consumption totals published in the Spanish IO table by applying a GME procedure. This estimation allows for identifying a considerable heterogeneity on the impacts derived of private consumption depending, not only on the administrative region, but also on the specific municipality within a given region.