The Poverty of Poverty Statistics

Hitler’s propaganda guru is reputed to have said that a lie repeated enough times eventually becomes a truth. Similarly, an “honest mistake” (assuming that there is such a thing as a “dishonest mistake”) that is left unchecked for very long eventually takes on the life of revealed truth. Such is the state of statistics on poverty in Ghana.

Sometime in 2000, someone somewhere credited the Fourth Ghana Living Standards Survey (GLSS4) with establishing that about 40 percent of Ghanaians (39.5 percent to be exact) lived below “the poverty line” in 1998/99, down from just over 50 percent (51.7 percent) in 1991/92.

Thereafter, GLSS4 became the oft-cited source of these and other poverty statistics in important policy documents, including the Ghana Poverty Reduction Strategy Paper (GPRS); the government’s budget statements; the president’s sessional addresses; and even policy analyses by some of the think tanks in the country. Government officials and NGO’s in their various public statements were no less enthusiastic in citing GLSS4 in matters of poverty in Ghana. More recently (May 2003), the voluminous and highly insightful Ghana Child Labour Survey also gave GLSS4 credit for Ghana’s poverty statistics.

The truth is GLSS4 says absolutely nothing about poverty levels in Ghana. As the name suggests, the entire 170 GLSS4 report deals not with poverty LEVELS but with living conditions in Ghana. In the report, we learn, for instance, that 53 percent of Ghanaian households live in mud huts (83.1 percent in the savannah); only 6.2 percent had “flush toilets” in their homes in 1999 (another 21 percent had none at all, while the others made do with KVIPs, pit latrines and pan/buckets), all in a country whose ruling elite routinely spend lavishsly on their personal comforts and leisure.

Incidentally, we also learn from GLSS4 that in 1998/99, Ghanaians received a total of C339 billion as remittances from abroad, which in March 1999 (the cut-off period for the survey) would be equivalent to US$142 million - or 40 percent of total household remittances in Ghana. (Even if this figure had doubled between then and now, it would not be anywhere near the $1.4 billion we are repeatedly told that Ghanaians abroad remit home annually).

For those interested in the moral dimensions of development and the complexities of the overall process of development, GLSS4 is a good starting point.

But while GLSS4’s contents implicitly reflect the conditions of poverty in Ghana, the report does not actually establish the number of people living in poverty in the country. That information comes to us from another publication, “Poverty Trends in Ghana in the 1990s,” which, like GLSS4, was produced by the Ghana Statistical Service in 2000.

Even then the Poverty Trends report is not without its flaws -- on its own and within the context of poverty reduction programmes. For one thing, the severe economic crisis that the country went through between mid-1999 and most of 2001 has effectively rendered some of the report's findings meaningless; the number of people living below the poverty line, in all likelihood, has increased significantly since the crisis.

But even if there had been no crisis, some of the conclusions and methodologies of the Poverty Trends report should have raised important policy questions then and now.

To begin with, the term “Trends” in the title is misleading, since two disparate data points at each end of a decade could hardly constitute a "trend," when no analysis is done of poverty in the intervening years. The suggestion by the report, therefore, that poverty showed a declining trend in the 1990s was misleading.

In 1995, in a reported titled “The Pattern of Poverty in Ghana, 1988-1992”, the Statistical Service reported that 31.4 percent of Ghanaians lived in poverty in 1991/92. Using a different methodology in 1998/99, however, the authors of the Trends report revised the 1991/92 figure upwards to 51.7 percent and then somehow determined that the new poverty line in 1998/99 was 39.5 percent, thus indicating a lower incidence of poverty in 1998/99 relative to 1991/92.

The new lines were based on “consumption poverty” of “C900,000 cedis per adult per year,” adjusted for inflation in 1999 prices. (Other measures of poverty, which were not used in the report but are significant from a policy perspective, are “income poverty” and “expenditure poverty”; each of course has its own strengths and weaknesses).

Leaving aside the technical details of how the new poverty line was derived, with a quick look at the underlying economic statistics and a little help from economic theory, it becomes highly implausible that the incidence of poverty in 1998/99 would be lower than it was nearly a decade earlier.

This is so when one considers the relatively weak economic performance of the latter half of the 1990s and in particular the damage that the energy crisis of 1998 did to the economy that year and to household income then and afterwards.

In 1991/92, for example, per capita income in Ghana averaged US$410; by 1998/99, this had declined to US$390. Theoretically, it is possible for the number of people living in poverty to decline even as per capita income falls, but this can happen only if there is a significant reduction in income inequality (or a massive redistributioin of income from the non-poor to the poor). This obviously did not happen in the 1990s or in 1998/99.

In fact, the measure of income inequality, known as the gini coefficient, rose to 60 in 1998/99 from about 34 in 1991/92. Since lower coefficients reflect more equal distributioins of income, we can safely conclude that inequality in 1998/99 in Ghana was worse than it was a decade earlier.

With a decline in average incomes and a rise in income inequality, there simply is no way poverty in 1998/99 could have been lower than it was in 1991/92. If anything at all, we can argue that the incidence of poverty in 1998/99 should have been at least 50 percent, and not 40 percent as widely believed.

These issues once again highlight the need for consistent and reliable statistics for economic management and national development planning. The importance of such consistency and reliability cannot be emphasized enough. Let our policy makers take note.

Views expressed by the author(s) do not necessarily reflect those of GhanaHomePage.