ИСТИНА

Value of statistical life (VSL) is a widely used instrument for risk monetizing in many countries, while in Russia, due to the lack of data, there are no credible estimates that could be comparable in terms of the methodology used. VSL estimates, representing numerically public preferences toward risks for human safety, are crucial for cost-benefit analysis corresponding to the government policies’ trade-offs on public health, education, demography, and environment. The most common way of obtaining values of statistical life is connected to individual preferences revealed in wage compensating deferential corresponding to a certain amount of fatal risk as a specific job characteristic [Viscusi & Masterman, 2017]. The key assumption is that workers who face a higher risk of death should be paid more than those who face a lower risk. Otherwise, they would prefer to switch from that job to a safer one [Hammit, 2000]. According to this, an empirical strategy represents estimating the hedonic wage function as a regression equation. At the next step, an estimated coefficient before fatal risk is used to calculate the value of statistical life itself, adjusting to the average wage rate and 100% risk of death [Viscusi, 2003]. A similar approach to the estimation of VSL in Russia is suggested based on microdata analysis, corresponding to the revealed preferences assumption. The basic idea is to determine the willingness of employees to accept compensation for a certain value of fatal risk within different industries corresponding to their job. Calculations are carried out using RLMS-HSE survey data and Rosstat data on fatal risks by industries. Using panel data from 2010 to 2020 gives the advantage of obtaining more accurate estimates compared to cross-section ones [Kniesner et al., 2012]. For this purpose, several econometric techniques are used, such as pooling regression estimation, between-group estimator, random and fixed twoways effects models. The corresponding estimation results are presented in Table 1. Even though at some point a fixed twoways effects model would be sufficient to get the most precise estimates, models of this type suffer from a severe sample shrinkage since only “switchers” who moved from one industry to another (more or less safe in terms of the fatal risk) within the given period contribute to the risk variation. For this reason, a random-effects model may be seen as a credible alternative [Kniesner et al., 2012]. Pooling regression and between estimates are provided for comparison since panel data structure is not considered. There were a few such works for Russia before, but they were based on very different methodological approaches. The current value of statistical life in Russia based on survey results in these works ranges from about 2.4 to 13.3 million rubles. At the same time, for instance, the amount of compensation for relatives of people who died because of a terrorist attack is fixed by law and equals about 1 million rubles; if people died on duty in the military forces, the amount of compensation would be 3 million rubles. For the comparison, according to the U.S. Department of Health and Human Services, the average VSL in the USA was 11.6 million dollars or about 836.8 million rubles. This striking contrast might partly be explained by a disparity in average wage rates and risk structure between Russia and the USA, but at most, it evokes from different methodological approaches. The resulting estimates of VSL in Russia using panel microdata are in the range from 231.7m (FE-model) to 420.6m rubles (between-group estimator), which is comparable to the corresponding values for the United States, considering the difference in GDP at purchasing power parity and more than 17 times higher than the estimates obtained earlier for Russia using the survey method, which is less credible due to hypothetical bias of stated preferences [Murphy et al., 2005]. These results can serve as an impetus for the start of a discussion on revising the VSL in Russia. Accordingly, the revision of the estimates should entail a more comprehensive understanding of the costs and benefits of various policy programs that explicitly or implicitly involve the value of human life and health.