Beyond proxy indicators: from indirect to direct measures of the underground sector in East-Central Europe.

• Author: Williams, Colin C.
• Position: Report

Abstract

To evaluate the size of the underground sector, numerous measurement methods have been employed ranging from indirect to direct survey approaches. Evaluating critically the range of techniques available, this paper firstly highlights the growing appreciation that direct rather than indirect measurement methods are more appropriate, reliable and accurate. Following this, attention turns towards the results of a study that has employed such direct survey methods to analyze the underground sector in East-Central Europe, namely the New Democracies Barometer (NDB) Survey. Examining eleven countries in the years 1992, 1994, 1996 and 1998, this reveals not only that underground work is either the most important or second most important contributor to living standards for around a quarter of all households but also that this sphere became more rather than less important to households in the decade after the collapse of the socialist bloc. Moreover, and contrary to the narrative that underground work is a survival practice for those excluded from the formal sphere, it is here shown that households relying on underground work for their livelihoods cope just as well, if not better, measured in terms of their ability to acquire consumer goods, as those reliant on the formal economy in East-Central Europe.

Introduction

The aim of this paper is to evaluate critically the use of indirect proxy indicators to evaluate the underground sector and to display how a consensus has begun to emerge that direct survey methods are more reliable, accurate and appropriate. Given that most of the direct surveys so far conducted have been small-scale qualitative studies, and that this has prevented cross-national comparisons being made, this paper reports the results of one of the only direct surveys to have produced comparative cross-national data on underground work.

Firstly, therefore, this paper will evaluate critically the various measurement methods that have been employed to analyze the underground sector. Revealing that the emerging consensus is that direct survey methods are more reliable and appropriate than indirect methods, attention then turns to reporting the results of one of the few cross-national direct surveys so far undertaken, namely the New Democracies Barometer (NDB) Survey, that compares the importance of the underground sector to living standards in eleven countries in East-Central Europe. The outcome will be to provide not only a baseline assessment of the importance of the underground sector to the post-socialist societies of East-Central Europe but also to raise some important issues about its resultant impact on household living standards that indirect survey methods have so far failed to identify.

Before commencing, it is important to define the underground sector. Similar to most other papers in this symposium, the underground sector is here defined in the now widely accepted manner as involving the paid production and sale of goods and services that are unregistered by, or hidden from the state for tax and/or welfare purposes but which are legal in all other respects (Feige, 1999; Portes, 1994; Thomas, 1992; Williams and Windebank, 1998). As such, only paid work that is illegal because of its non-declaration to the state for tax and/or social security purposes is included under the umbrella of the underground sector. Monetised exchanges where the goods and/or services are illegal (e.g., drug trafficking) are excluded, as are non-monetised exchanges.

Measurement methods: a critical evaluation

Given that the underground sector is by its very nature hidden from, or unregistered by, the state, estimating its size is a perplexing and difficult task. The methods so far used range from techniques that employ direct survey methods to those that attempt to indirectly measure its size by seeking statistical traces of the underground sector in data collected for other purposes (see Bajada, 2002; Thomas, 1992; OECD, 2002; Renooy et al, 2004; Williams, 2004a; Williams and Windebank, 1998). These indirect methods use various measures ranging from non-monetary proxy indicators (e.g., based on labor force estimates, the number of very small enterprises, electricity demand), through monetary indicators (e.g., of the number of large denomination bank notes in circulation, the cash-deposit ratio or level of monetary transactions) to income/expenditure discrepancies either at the household and/or national level.

Indeed, the major reason that indirect evidence is sought in macroeconomic data collected and/or constructed for other purposes is because there is a pervasive assumption that respondents directly approached will not be forthcoming about whether or not they engage in underground work. For those who use proxy indirect measures, therefore, the belief is that even though these underground workers wish to hide their incomes, their work will be nonetheless revealed at the macroeconomic level so it is through these statistical traces that the level of underground work can be identified. At the other end of the spectrum are those who assume that despite the illicit nature of underground work, reliable data can be collected from research participants. These analysts thus conduct direct surveys. Here, a critical review is provided of measurement methods used to date so as to provide an overview of how data is collected and the problems involved.

Indirect methods

Indirect methods are of three kinds. Firstly, there are those that seek statistical traces of such work in nonmonetary proxy indicators such as discrepancies in labor supply figures or the number of very small firms. Secondly, there are those that track down evidence of underground work in monetary indicators collected for other purposes, and third and finally, there are those that investigate discrepancies between income and expenditure levels either at the household and/or national level. Each is here considered in turn. The problem with all of these methods, as will be shown, is that these indicators not only provide an unreliable and inaccurate proxy of the volume of underground work but also are unable to provide any understanding of the nature or distribution of such work. Instead, in most cases, they make some very crude assumptions concerning its character that are far from proven.

Indirect non-monetary indicators

The two most common indirect approaches that use non-monetary surrogate indicators to estimate the extent of underground work are firstly, those that seek traces of underground work in formal labor force statistics and secondly, those that use very small enterprises (VSEs) as a proxy for the existence of underground work.

Labor force estimates. Methods that seek to measure underground work from formal labor force statistics are of two types. Firstly, there are those that identify various types of employment (e.g., self-employment, second-job holding) as proxy indicators of the existence of underground work and then look for unaccountable increases in official labor force statistical data on the numbers employed in these categories (e.g., Crnkovic-Pozaic, 1999). The problem, however, is that the idea that underground work prevails in these categories of employment is an assumption, rather than a finding, of the technique, and there is no way of discerning the extent to which it is underground work, rather than other factors, that has led to an increase in these categories of employment. The growth of self-employment, for instance, is not simply a result of the rise in underground work. It is also due to the rise of an enterprise culture, increased sub-contracting in the production process and the advent of other forms of flexible production arrangements. Second-job holding, furthermore, is by no means a direct result of underground work except if such job holding is illegal per se. It is also in part the combined result of broader economic and cultural restructuring. To identify the proportion of growth in either self-employment or multiple jobholding attributable to such processes and the share attributable to underground work at a particular moment is thus a difficult if not impossible task.

Secondly, therefore, those who analyze official employment data for evidence of underground work have sought discrepancies in the results of different surveys used to compile official employment statistics (e.g., Denison, 1982). However, many problems exist with this approach. The first problem as both Bajada (2002) and Williams and Windebank (1998) highlight, is that it erroneously assumes that each individual is either a licit or illicit worker and in so doing, misses the vast majority of underground work conducted by those who have a formal job (Williams, 2004a). Secondly, by analyzing only those employed in businesses, it assumes that all underground work is conducted on an 'organized' basis and misses both more autonomous forms of underground work and underground work conducted for households. Thirdly, there is no reason to assume that an underground worker will describe him/herself as employed in a household survey whilst the employer will not in a business survey. And fourth and finally, the fact that such analyses have resulted in contradictory results, with some studies showing no change in size of the underground sector in the post-war years (e.g., Denison 1982) and other showing growth (e.g. US Joint Economic Committee, 1983) intimates the need for great caution. Identifying the magnitude of underground work through formal labor force statistics, in sum, is beset by problems that cannot be easily transcended and this method has waned in popularity since its heyday in the early 1980s. Indeed, Eurostat in a note to the informal Council of Ministers of Employment and Social affairs on undeclared work, conclude that 'a comparison of the labor force survey results with administrative sources or register data on an aggregate level does...