Consumption smoothing and vulnerability in Russia

Abstract

Applying bootstrapped quantile regression to the Russian Longitudinal Monitoring Survey (RLMS) data, we examine the channels through which individuals experience and seek to cope with changes in consumption. We find that married individuals living in small households, with educated heads in urban areas are better equipped to smooth consumption. Investigating the impact of idiosyncratic shocks, we find that the labour market is an important transmission mechanism allowing households to smooth their consumption but also exposing them to risk, mainly through job loss. Outside of pension payments, the formal social safety net does not facilitate consumption smoothing, thus heightening the importance of informal coping institutions. It transpires that both support from relatives/friends and home production act as important insurance mechanisms for the most vulnerable. In contrast with previous findings, it would seem that regardless of its historical, political and social roots, the garden plots and dachas, often romanticized in Russian literature, do provide a means by which ‘urban’ Russians are able to cope with economic fluctuations. We finish by stressing the important policy lessons for Russia's developing market economy.

Notes

¹ See, for example, Klugman (1997), Ovcharova et al. (1998), Braithwaite (1999) and Falkingham and Kanji (2001).

² See Bradbury et al. (2001) for a discussion regarding the use of the individual or the household as the unit of analysis.

³ The SEs for the quantile regressions are bootstrapped, bias corrected and based on 100 repetitions for data clustered by household.

⁴ Apotential weakness of the RLMS consumption data is that it is based on recall over the previous week or month rather than on records kept. However, not only are the consumption levels consistent over time relative to reported income, but the information is collected at a similar time of year in each round and so, though we add a note of caution in interpreting our results, we have no reason to believe that any data deficiency is nonrandom.

⁵ We have also tried with other equivalence scales and without equivalizing and find little qualitative change in our key results.

⁶ We make no attempt to measure the level of accumulated wage arrears since the data does not identify exactly when the arrears accrued and therefore it is not possible to identify its value in real terms. Instead we combine it with ‘unpaid leave’ to try to capture informal labour market mechanisms resulting in ‘shocks’ to income.

⁷ We additionally ran the estimates with ‘initial levels’ of the coping variables and find that all results hold.

⁸ Even in urban and metropolitan areas agricultural production for personal consumption (or sale) has always formed a significant part of the resources of Russian households. See Seeth et al. (1998), Clarke (2002) and Pallot and Nefedova (2003) for a more detailed discussion.

⁹ In view of the well-documented phenomenon of multiple job holding, we experimented with an ‘additional job’ variable as a coping mechanism but found it not to be significant. We also broke down the ‘work’ variable into individual occupations and though other results remained robust to this alternative, we felt that the story was really one of working or not rather than one of particular occupational affiliation. In addition, our procedure removes the problems associated with individuals changing occupation in response to the economic climate.

¹⁰ Practical estimation and inference methods for instrumental quantile regressions are complex and subject matters of recent research. See, for example, Chernozhukov and Hansen (2006).

¹¹ For ease of exposition, the results of the regressions are not reported in this article but are available upon request.

¹² See Gerry et al. (2008) for a more detailed discussion of nonrural advantages.