ity of individuals increased. The changes in the relative income of two vulnerable groups, the young and the old, were mixed. There was improvement for both groups in the Czech Republic, improvement for children in Hungary, and improvement for the elderly in Poland. But given their considerable limitations, the data can only be suggestive of income changes; by themselves they cannot be used conclusively.
The sources of bias in the survey data have been mentioned above. The surveys exclude a small but important segment of the population—the institutionalized, the military, and the police. Since 1989, the surveys also have had an increasing nonresponse rate. The effect of this increasing trend on the data is not yet known. Most important, the full effect of the informal economy is not captured in any of the Central European surveys; as income from this economy grows, its exclusion from the surveys will become increasingly important. Without more research, the magnitude of these biases cannot be accurately estimated.
Adding to the known biases of the income survey data is the incompleteness of the concept of income for measuring welfare (van de Walle, 1996). While income is an important factor in welfare, other concepts, such as the measurement of time allocation, life expectancy, and the subjective sense of personal well-being, are very important. Private transfers of time and resources among social and family networks may also be very important in the transition period.
The governments of Central Europe need reliable household income data to help target social safety net transfers to the most vulnerable individuals and households. In both the Czech Republic and Poland, social safety net programs have been used explicitly to minimize the adverse effects of the economic transition on pensioners (Blanchard et al., 1994). The sustainability of current pension levels is, however, in doubt (Sachs, 1995). If social safety nets decline in the future, the need for reliable household income data to guide the allocation of shrinking resources will become even more important.
The Central European countries studied in this chapter have had a long tradition of taking sophisticated household income surveys. The economic transition, however, has made the quality and quantity of the data an issue. More work needs to be done to improve the quality of the data, and more resources will be needed to improve the quantity and timeliness of the surveys. Future research should also focus on the many indicators of well-being that are not captured by income measures, no matter how accurate they may be.
The current data raise a number of intriguing questions. For instance, how closely linked are the economic transitions of the macro, national economies and the micro household economies? What adjustments do households make to insulate themselves from major declines in the macro economies? Do households pool their resources, and if so, how does this happen? Are the