Sat, 27 Sep 2003

Politics of numbers show economic growth

Peter Gardiner and Mayling Oey-Gardiner, Researchers, Insan Hitawasana Sejahtera, Jakarta

After the severe economic slump of 1998 and 1999, Indonesia has returned to a period of significant positive economic growth. Not as rapid as during the pre-crisis period, but still not that far off regional averages and, of course, considerably better than in many other parts of the world.

But is this fast enough? Not according to many economic gurus, who have predicted a bleak future for the labor force in terms of rising unemployment and underemployment unless growth returns to something close to pre-crisis levels. The basis for the grim predictions lies in a selective reading of some of the labor force numbers coming from the National Labor Force Surveys (Sakernas) as published by the Central Bureau of Statistics (BPS) and assumptions on the relation between net additions of workers and economic growth.

The core of the argument apparently lies in the last three years of the labor force statistics that suggest a significant rise in unemployment from about 6 million in 2000 to 8 million in 2001 and just over 9 million in 2002 (see Table 1). These are equivalent to unemployment rates of 6.1, 8.1 and 9.1 percent respectively. On this basis (and using current economic growth assumptions) it is suggested that unemployment could grow to 10 million in 2003 for an unemployment rate of just under 10 percent.

Given economic growth projections of about 4.5 and 5.0 percent in 2004 and 2005, and the rather interesting assumption that each percent of growth means 350,000 net new jobs, means that during the next two years only about 3.3 million people would be added to the current workforce. But the labor force is projected to grow by some 4.2 million people. Hence, further increases in unemployment to the tune of close to 1 million additional people out of work.

The basis for these predictions, however, is questionable. We need to ask why these gurus looked at so few numbers and did not seek to put them in a broader context or ask serious questions about what they actually mean. Is the medium the message? Is what is being said more provocative than real?

There are some significant problems with the BPS data. For example, we need to be aware that the labor force data are based on a sample survey and the sample has to be weighted to totals obtained from population projections to arrive at the final absolute numbers. But population projections are just that-- projections, and periodic adjustments can result in discontinuities that are hardly a reflection of reality.

Just look at the recorded growth of the working age population (15+) between 1999 and 2000 in Table 1. That was solely due to the adoption of a new base in 2000 from the population census.

In fact, the entire time series on the population age 15+ shows erratic fluctuations with the size of the working age population increasing by 3.2 million between 1996 and 1997, 3.5 million between 1997 and 1998 and only 2.5 million between 1998 and 1999. The stagnation in growth between 1999 and 2000 was, as noted above, a result of adjustments to agree with new totals from the 2000 census.

But even after the census, there are questionable trends-- growth of 2.9 million between 2000 and 2001 and then fully 4.7 million between 2001 and 2002. Given that growth in the working age population should be slowing as a result of past fertility decline, this latter figure is particularly problematic.

Numbers in the labor force, numbers working and numbers unemployed also fluctuate according to these data and not always as might be expected. With unemployment there is also the problem of a change in definition in 2001 that added independently to the estimated number of unemployed. Thus between 1996 and 2000, the survey instruments remained practically the same and unemployment grew from around 5 percent in 1996 to about 6 percent in 2000 with a moderate (but hardly overwhelming) upsurge in 1998 and 1999 due to, as would be expected, the crisis.

The change in definition in 2001, however, resulted in an increase in the number of unemployed from 5.8 to 8.0 million. And a third of this total (2.7 million) was actually accounted for by the definitional change. Keeping the new definition, the 2002 Sakernas estimated unemployment at 9.1 million of whom 3.2 million, or 35 percent, represented the impact of the change in definition.

In other words, if a constant (pre-2001) definition had been used throughout, the numbers of unemployed would have only been 5.3 and 5.9 million respectively in 2001 and 2002 and actually suggesting an improvement since 1999, running directly counter to the "doom and gloom" prognosis of the economic pundits.

Furthermore, any clear relation between economic growth and job creation is hard to see from these data. For example, between 1996 and 1997 there was an increase in the number of jobholders of some 1.5 million while the economy grew at 4.7 percent. But then, between 1997 and 1998, practically at the height of the crisis, the number of workers continued to increase by some 2.3 million at the same time that the economy was imploding with a contraction of 13.1 percent.

And in the following years increases in the working population have hovered at around 1 million per year while there has been substantial fluctuation in economic growth. Where the "magic" figure of 350,000 new jobs for each percent of economic growth comes from is hard to determine, at least from these numbers.

The gurus also make a lot out of the number of underemployed -- defined as those working less than 35 hours a week and who constitute around 30 million people or one-third of the workforce. But who are they in reality? Most (about 70 percent) of these underemployed are in agriculture and most of these are small-scale self-employed farmers, fishermen, etc. Not surprisingly, therefore, they tend to be relatively poorer and have had less schooling -- around three-fourths have at most an elementary education.

But concepts of underemployment that really relate better to non-agricultural wage labor have less meaning here where labor inputs are determined more by other factors such as season and weather. And this says nothing about the difficulties of effectively recording working time across a range of different jobs that many of these people do.

In fact, the real problem with labor markets arguably may have less to do with formal concepts of unemployment or underemployment than with the nature of the work that people do, particularly their employment status. Here a major distinction lies between those with access to, presumably more regular, wage or salary employment and those who are self-employed or working as unpaid family workers in what is commonly referred to as the non-formal sector.

The formal sector in Indonesia remains relatively small with about one-third classified as either employers (ca. 3 percent) or employees (ca. 30 percent). Two-thirds of the workforce is in the non-formal sector as own-account workers, self-employed, casual workers and unpaid workers. In this non-formal labor market, own- account workers and the self-employed constitute the majority at 42 percent, with the remaining 25 percent made up by unpaid family workers.

This last group is of particular interest. According to the latest Sakernas of 2002, unpaid family workers numbered some 16 million people, much higher than the number of unemployed. Yet these people are actually working (and often quite hard) for no immediate financial gain.

Why are they less of a concern? To the pundits this group seems apparently invisible. Is it because these unpaid workers are overwhelmingly female (at three-fourths of the total), while the pundits are predominantly male?