Most of the empirical work on lateral pressure theory address the propensity for expansion of behavior outside territorial boundaries with reference to actual behavior (rather than propensity). While this is entirely consistent with the theory, it bypasses the thorny problem of metricizing the propensity variable and then examines its connections to actual behavior. More recently we developed the Lateral Pressure Index in order to quantify propensity for expansion and to the extent possible, to highlight the relative salience of individual drivers. After some experimental, we framed the Lateral Pressure Index as a function of the geographic mean of its master variables:
An alternative choice for constructing Lateral Pressure Index could have been to use linear aggregation. Once marginal contribution of each variable are separately assessed, such contributions can be added to calculate the cumulative yield. Using linear aggregation technique requires a necessary and sufficient condition that the indicators are mutually and preferentially independent. Further, the use of linear aggregation may result in development of a biased indicator that does not entirely reflect the information of individual indicators. OECD (2008: 103).
Any composite indicators based on the additive aggregation show an undesirable feature of implied full compensability, i.e. a poor score in one dimension can be compensated for by sufficiently high score in other dimensions. For example, if the Lateral Pressure index, formed by population, resource and technology index for two states, one with values 0.15, 0.05, 0.01 and the other being 0.06, 0.07, 0.08, will have the same composite Lateral Pressure Index score if linear aggregation with equal weights are applied. The composite Lateral Pressure Index based on linear aggregation would not be able to differentiate two nations with totally different state profiles. Use of geometric aggregation helps in avoiding this undesirable feature. In the example above, if the aggregation were geometric the aggregated sores will be different (0.0422 for the first and 0.0695 for the second).