The Social Impact of the Indonesian Economic Crisis Labour by han19590


									                                        EADN RP1-2

EADN Regional Project on the Social Impact of the

             Asian Financial Crisis

The Social Impact of the Indonesian Economic

     Crisis: Labour Market Adjustments


         Tubagus Feridhanusetyawan

                  January 2002

                      LABOUR MARKET ADJUSTMENTS

                                   Tubagus Feridhanusetyawan

Indonesia’s economic development for the three decades before the crisis was impressive in
terms of macroeconomic indicators. The economy grew by an average of 6 per cent annually,
inflation was low (mostly in single digits), and the current account deficit was manageable,
supported by strong export growth. Per capita income reached the $1,000 threshold in 1996,
and the proportion of the population in poverty declined from 40 per cent to 11 per cent in the
last 20 years. Even in mid-1997, when Thailand was already in deep trouble and only a few
months before Indonesia floated the Rupiah, the World Bank and the International Monetary
Fund still saw Indonesia as an example of a success story of economic development.

However, from mid-1997, Indonesia was in total crisis – economic, social, and political –
which prevailed till as late as 1999. Economic contraction continued, and the process of
economic restructuring and adjustment progressed very slowly. It was clear that a crisis of
confidence had played a major role in the deteriorating performance of the Indonesian
economy since the crisis started in mid-1997. This first emerged in late 1997 when there was
no clear political will by the government to conduct consistent reforms. The political
misfortunes led to a situation where the economy, and even the country as a whole, lost the
mechanism for self-correction. This political uncertainty lingered throughout 1998 and 1999,
which clearly prolonged the process of economic recovery. The collapse of the economy led
to various social and economic problems in Indonesia both in the short and long run, ranging
from increased unemployment, declining real income, and a lower quality of social services.

There was some initial controversy over the magnitude of the social impact of the crisis,
especially regarding unemployment and poverty. As early as the latter half of 1997, many labour
economists had predicted that due to labour market flexibility, the unemployment rate would not
be high and could not be used to represent the distressed labour market. It had been predicted
since the beginning of the crisis that a massive adjustment would take place in the form of
declining real wages. Despite this, labour union activists, politicians and even economic
observers in the media (some of them are economists) had quoted unemployment figures that
were three to four times larger than the real and actual unemployment rates calculated from
subsequent surveys. This large divergence may have been partly because of not distinguishing
between the increasing number of layoffs (which was high) and the increasing number of the
unemployed (which was small).

Another controversy was regarding the proportion of the population living below the poverty line
during the crisis. The estimate ranged from 48 per cent by the International Labour Organization
(ILO) to 14 per cent by the World Bank, based on the different methodologies for measurement
using various factors such as price deflators, and different data. People could clearly feel the
severity of the social impact of the crisis in their daily lives, but capturing the social impact of the
crisis and measuring the lower quality of life of the Indonesian people after the crisis turned out
to be a complex task.

One objective of this paper, therefore, is to place the social impact of the crisis in the right

perspective by examining adjustments in the labour market. The social impact of the crisis covers
a wide range of areas, from the short run effects such as adjustments in the labour market and
poverty issues, to longer run problems related to health and education. This paper focuses mostly
on the short run aspects, and in particular on the various adjustments in employment,
unemployment, income and other issues related to the labour market. It is shown that the theory
and the empirical evidence justified lower unemployment rates. Although the labour market had
been adjusting very well during the crisis, the longer-term social impact of the crisis in the form
of lower quality of education and health related problems could be more serious. In order to
survive in the short run, families had been substituting their expenditure on secondary items such
as children’s education and health for more basic items such as food. Therefore, future policy
should also take into account the longer -term impact on education and health.

The paper starts by exploring how the economic crisis was transmitted into the social sphere, and
then reviews how this transmission mechanism worked in particular relation to the labour market,
employment and wages. The second section will then examine more specifically the changing
structure of the labour market, including employment, unemployment, and real income. The third
section focuses on the emerging trends in the labour market such as the reversal process of labour
market transformation, labour migration, and the increasing number of women and children in the
labour market. The fourth section presents various issues related to the distributional impact of
the crisis, including poverty issues. Finally, the fifth section presents the summary, and some
policy recommendations. In view of the limited scope of the paper, specific policy responses,
such as the design, implementation and shortcomings of social safety nets are not discussed.


A chronological review of the events leading to the crisis suggests that Indonesia’s crisis was
caused by a series of mistakes, both in terms of policy, and in terms of institutional and legal
failures, which turned the initial financial crisis into a nationwide economic crisis.

Prior to the crisis, in early 1997, Indonesia’s macroeconomic fundamentals were still strong, and
Indonesia was considered in a better position to face the crisis compared with Thailand. 1 From
August to December 1997, the crisis in Indonesia followed a path similar to that in neighbouring
countries such as Thailand and Korea. Exchange rates depreciated by around 50 per cent, the
money supply was under control, and inflation was still in single digits. In fact, the overall
economy still grew at around 4 to 5 per cent throughout 1997. 2

Between December 1997 and January 1998, however, Indonesia moved to a different and
worsening path. By the end of 1997, there was a drastic reversal of market expectations of the
Indonesian economy, and especially towards the commitment and credibility of Soeharto’s
administration in handling the crisis. It became clear that the International Monetary Fund’s
(IMF’s) first package in November 1997 was not fully implemented by the government, and it
was obvious that without any significant political reforms, it would be difficult to implement
economic reforms. Combined with the IMF-approved liquidation of the 16 banks with no
provision of deposit guarantees in place, confidence in the government, the banking system and
the economy in general was low. The deteriorating confidence in the economy eventually led to a
massive bank run, capital flight, and an increase in expected inflation.

The massive bank run that took place from December 1997 until January 1998 led to a large
monetary expansion and plunging exchange rates. The political crisis throughout 1998 worsened
the situation. The riots in May 1998 followed by the change in national leadership created further
chaos in the economy, another bank run, and further depreciation of the exchange rates. In
response to the weakening exchange rates, interest rates increased up to 60-70 per cent at the peak
of the crisis in mid-1998.

From crisis of confidence to economic collapse
The financial crisis was compounded by what could be termed as a crisis of confidence (triggered
in part by the lack of clear political direction of the government in responding to the crisis) which
emerged by late 1997. Figure 1.1. illustrates the transmission from the crisis of confidence in late
1997 to the collapse of the economy in mid-1998.

 The officials from the World Bank and the IMF made several positive statements about the Indonesian
economy in July 1997, a month before the Rupiah was finally floated.
  For a discussion on the transmission from strong economic growth as a result of economic deregulation in
the late 1980s and the causes of the crisis and the deteriorating confidence in the economy, see, for
example, Feridhanusetyawan (1998).

Figure 1.1. From Crisis of Confidence to Economic Collapse

                                         Crisis of Confidence

   Weak Banking

                          Bank Run                        Capital Flight

  Monetary Expansion                                               Credit Crunch
  To Bail-out Banks

              High Interest
              Rate Policy

                              Banking                                Insolvency

                                                                           Real Sector

  High                   Rupiah                            Falling Output and              Falling
  Inflation              Depreciation
                                                              Employment                 Government

                                            Falling Real Income

The loss of confidence in the economy in late 1997 led to both a bank run and capital flight. In
late January, the massive bank run and excessive monetary expansion forced the government to
provide a blanket guarantee to all deposits in all amounts, in domestic banks, to prevent the
collapse of the banking system. At the same time, the provision of liquidity support for failing
banks led to an increase in the currency in circulation, especially when the provision of liquidity
support was partly financed by printing new money.

Excessive monetary expansion further fuelled rupiah depreciation and high inflation. To curtail
excessive monetary expansion, the interest rate was kept high. However, combined with high
inflation, real interest rates became in essence negative in the first few months of 1998, causing a
strong disincentive to hold Rupiah deposits. There was a dramatic jump of money in circulation
by 32 per cent in January, and 15 per cent in May 1998. In the first six months of 1998, the
money in circulation increased by around 60 per cent (a 100 per cent growth from November

1997). Despite the fact that Indonesia was on an IMF programme, reserve money growth
continued to be high at 47 per cent, indicating that excessive liquidity remained in the economy
for some time.     When renegotiating IMF reforms after President Soeharto was elected again in
the first week of March 1998, it was agreed that interest rates needed to be raised to absorb the
excess liquidity created in the first few months of the year. SBI (Bank Indonesia certificates)
rates were doubled in the third week of March to 45 per cent.

High interest rates in turn, led to further deterioration of banks’ balance sheets, and furthermore
led to corporate and economic insolvency. The dysfunctional financial sector eventually led to a
fall in real sectors, resulting in falling output, employment and government revenues. A high
interest rate environment also hurt the stock market, with a decline in both stock prices and the
volume of transaction. The volume of transaction dropped from around US$5 billion in July
1997 to slightly more than US$100 million in August 1998. The Jakarta Composite Stock Price
index decreased by around 50 per cent, from 721 in July 1997 to around 350 in August 1998.
The stock price index remained at around 400 in early 1999.

The total capital outflow was unrecorded as there was no reporting requirement for any foreign
exchange transaction. Nonetheless, from the capital account in the balance of payment it is clear
that there was a large net outflow of capital, with a reversal in private net capital flows, from
inflows of $13.5 billion in 1996/97 to net outflows of $12 billion in 1997/98. Another estimate,
from the World Bank, showed that the capital outflow was up to around 15 per cent of GDP.3
However, this may be an underestimate of the total capital flight because the flow of capital in
and out of the country was unrecorded. This massive capital flow clearly led, among other
factors, to a sharp depreciation of the rupiah.

The exchange rate depreciated from Rp. 2,400 per US dollar before the crisis in July 1997, to Rp.
15,000 per US dollar in July 1998, before reverting to around Rp. 9,000 per US dollar in
September 1998. The depreciation of the exchange rate (up to 600 per cent at its steepest) was
mainly caused by excessive monetary expansion. Due to the excessive printing of money in
January and February 1998, and again in May 1998 after the riots, the depreciation of the Rupiah
dramatically deviated from the patterns of the other Southeast Asian currencies.

While Indonesia shared similar problems with other Asian countries in crisis, such as the credit
crunch and a huge foreign debt, the political turmoil combined with explosive monetary
expansion, high inflation, high interest rates and massive rupiah depreciation were the special
factors that set Indonesia apart from the other Asian countries in crisis. As a consequence,
Indonesia’s real sectors were also the worst affected in Asia.

The economy went into a deep recession in 1998, with overall negative growth rates. The worst
contraction was in the construction sector (-39.8 per cent), financial sector (-26.7 per cent), and
the trade, hotel and restaurant sector (- 18.9 per cent). Other sectors that experienced large
contractions were manufacturing (-12.9 per cent) and transport and communications (-12.8 per
cent), while mining and other services sectors contracted by around 4.5 per cent. On the other
hand, the agricultural and the utility sectors still experienced positive growth at around 0.2 and
3.7 per cent respectively. The share of agriculture in GDP rose for the first time in 30 years, from
16.1 per cent in 1997 to 18 per cent in 1998, a reversal of the trend of the structural
transformation of the economy.

Although the crisis began with a financial crisis in 1997, the collapse of the financial sector,

    World Bank (1998a).

marked by significant bank restructuring and asset liquidation, only took place at the end of 1998.
The complex problems related to bank restructuring, the absence of credible bankruptcy laws, and
other problems such as the lack of capacity within the government to deal with the problem,
further prolonged the needed and necessary adjustment in the financial sector.

 From the expenditure side of growth, the contraction was mostly caused by a severe decline in
investment. The decline in private consumption by around 2.9 per cent was smaller than the
decrease in government expenditure at more than 14 per cent, and investment at around 40 per
cent, which suggests that people were smoothing their consumption by dissaving during the
crisis. It was estimated that per capita national income (based on current prices) declined from
around US$1,055 in 1997 to around US$ 460 in 1998 (based on a 13 per cent economic
contraction, and an average exchange rate of Rp. 10,000/US$).

The effect of the economic crisis on inflation was dramatic. Prior to the crisis, inflation was
maintained below 10 per cent. However, by the end of 1998, cumulative inflation was around
77.6 per cent. While the rise in food prices by more than 118 per cent dominated the increase in
the Consumer Price Index (CPI) in 1998, the rise in fuel and electricity prices (which raised
transportation costs) and the riots in May 1998, which temporarily destroyed the distribution
system, worsened the situation.

From economic collapse to social problems

There are three main channels by which the economic crisis was transmitted into a social crisis:
adjustment at the macro level of the output and input markets, especially the labour market;
adjustment at the micro level, namely the changing patterns of household income and
expenditure; and direct transmission through government expenditure, namely the provision of
public social services. Figure 1.2 presents these transmission mechanisms.

The first channel of transmission was through adjustments at the macro-level in the output and
input markets. On the output side, the crisis has moved resources from the modern, non-traded
and import-dependent sectors, such as construction and capital intensive manufacturing, to
traditional, traded and export-oriented sectors such as agriculture, forestry, mining and labour-
intensive manufacturing. Firms in the shrinking sectors have been severely affected through
market adjustments in the form of declining levels of profit and real income, firm insolvency and
company closures. On the other hand, some sectors, such as export-oriented manufacturing and
agriculture, had in fact benefited from the crisis. This has led to complexities in assessing the
social impact of the crisis. Data suggests that the social impact of the crisis was heterogeneous
across regions, gender, income groups, sectors and industries.

On the labour market side, the excessive monetary expansion and high inflation that characterized
the Indonesian crisis, combined with the flexibility of the labour market led to different
adjustments in the labour market compared with that in Thailand or Korea. In these two
countries, with relatively low inflation, and relatively stable real wages, adjustment was in the
form of rising unemployment rates. In Indonesia, the flexibility of the labour market and high
inflation rates led to a drastic fall in real wages and a small increase in the unemployment rate.4

  To compare the social impact of the financial crisis in various Asian countries, see Manning (1999), and
also Lee (1998).

Such macro-level adjustments in the labour market affected the translation of economic
contraction into falls in employment, real income and labour productivity. As a result of a
flexible labour market in a naturally labour surplus economy, the increases in unemployment and
underemployment rates were relatively smaller than previously expected, but the decrease in real
wages was drastic. Consistent with the changing structure of the economy in general, labour
moved from the formal to the informal sector, from modern to traditional sectors, and from urban
to rural areas. The discussion about the adjustments in the labour market will be presented in
more detail in the next section.

The second channel of transmission was at the micro level, in the form of changing patterns of
household income and expenditure. During the crisis, sharp reductions in real income forced
people to work more for relatively less income, to consume their savings, or sell their assets to
cope with increasing expenditure. The purchasing power of families also declined, as the increase
in prices was three to four times larger than the increase in nominal wages. The sharp reduction
in real family income also forced women and children to enter the labour market, and to seek
additional family income. The increasing number of street children and beggars in urban areas
revealed how vulnerable groups were the real victims of the crisis.

In relation to expenditure, the doubling of food prices forced people to reduce or substitute their
spending on secondary and tertiary needs such as clothing, health services, education and housing
for more basic needs, especially food. For low-income families, where food consumption
accounts for most of their expenditure, the sharp increase in food prices significantly reduced
their purchasing power, lowered their food consumption, and even led to starvation in some

The substitution from secondary or tertiary expenditures to more basic commodities such as food
could lead to serious human capital problems in the longer run. The decline in real income forced
many families to reduce their spending on children’s education, and even to withdraw their
children from school. Expenditure on health services, including contraceptives and immunization
was also reduced.

Another major channel through which the economic collapse was transmitted to social problems
was through direct government expenditure. The government’s budgetary constraints during the
crisis led to lower public spending on education, health and other social services. Combined with
the lack of capacity within the government to deal with the crisis, the limited government budget
resulted in ineffective social safety net programmes to mitigate the social impact of the crisis.

Along with these transmission mechanisms, the drought or the El Niño effect in 1997 magnified
the social problems in Indonesia during the crisis. In addition to the breakdown of the distribution
system after the riots in May, the shortage of domestic food supply due to the drought in 1997
contributed to high food price inflation. The price of food increased 50 per cent more than the
general inflation rate, constraining purchasing power especially for the poor, and raising the
potential for food security problems.

The problems associated with the social impact of the crisis were tremendous in the context of the
difficult political and other reform processes in Indonesia throughout 1998-1999. The
combination of the break-down of law and order, the loss of government credibility, the sharp
decline in real income and the unavailability of social safety nets led to various riots, social unrest
and serious ethnic conflicts in various places in Indonesia.

The IMF-led reforms and employment

While the economic crisis was transmitted to the social sector through these various channels,
other factors, such as the IMF-led economic reforms following the crisis also had a significant
impact, both in terms of structural reallocation of economic resources and welfare implications.
The reforms, which were formulated under the agreement with the IMF, consisted of three broad
categories: macroeconomic policies, financial sector restructuring, and structural reforms. It is
difficult to distinguish the impact of those economic reform policies on employment and incomes,
because the impact of the economic crisis itself and the policy response to the crisis cannot be
separated. This is further complicated by the fact that the reform policies had been revised several
times since the crisis in response to the rapidly changing situation. The progress of the reform
process was also slow because of the unfavourable political climate, even with the new

While recognizing the difficulties in analysing the impact of the reforms on employment and
income, it is generally believed that in the long run everyone benefits from reforms because of
more efficient resource allocation and higher economic growth. Such economic adjustment
entails the expansion of some sectors and the contraction of others. In the short run, therefore,
workers in the contracting sectors and other vulnerable groups suffer, and hence, there is a need
for social safety nets to offset this pain of economic adjustment.

Structural adjustment programmes, which mainly consist of the removal of existing import
barriers, domestic distortion in the form of monopoly practices, and restrictions on inter-
provincial trades, generally tend to pave the way for a more competitive economy. This however,
hurts those sectors and firms, which were previously protected and inefficient. Similarly,
financial sector restructuring reduces distortions in the banking sector, but the consequent
downsizing and merging of banks leads to unemployment among white-collar workers.

Macroeconomic policies, especially high interest rates and tight monetary policy, generally tend
to have differential effects. It prevents the inflation rate from catching up with the depreciation
rate, which means that fixed salary earners tend to benefit from the policy in terms of less
deterioration of their purchasing power. However, small businesses and industries that do not
depend on foreign input are hurt by the high interest rate environment, which increases their cost
of capital. As part of fiscal policy, the removal of government subsidies, especially on fuel and
electricity, and an increase in tax, for example in alcohol and tobacco, leads to a higher price of
transportation and utilities. In the Indonesian context, the removal of fuel and electricity
subsidies was however, delayed after the May riots.

   Figure 1.2. From Economic Collapse to Social Problems

                              Falling Government                  Rupiah                       High                 Falling Output and
                                    Revenue                     Depreciation                 Inflation                 Employment

                                 Adjustment in Household (HH)                                    Adjustment in the Labour          Adjustment in the Output
                                    Income and Expenditure                                        Market: Labour Surplus         Market: Changing structure of
                                                                                                       Economy and                       the economy
                                                                                                  Flexible Labour Market

                 Adjustment in HH Expenditure             Adjustment in HH Income                        Small increase in      Large Reduction
                 -Sharp reduction in purchasing           -Work more for less salary                     Unemployment &         In Real Wage &
                 power.                                   - Selling Assets                               Underemployment        Productivity
                 -Smaller saving or dissaving             - Forcing family member
                 -Lower spending on education              (women and children) to
                 -Lower spending on health                  work                                         Changing structure of the labour market
                 -Lower quality and quantity of                                                          -  From formal to informal sector
                  Consumption – especially food                                                          -  From modern to traditional sector
                                                                                                         -  From non-traded to traded sector
                                                                                                         -  From import to export sector

                                                                                                         - Greater uncertainty in the labour market
El Nino                                                                                                  - Massive Layoffs in the formal sector

    Food              -Lower public         Health Related Problems            Education Related Problems           -Women and children          -Increase in
    Insecurity        spending on           - Lower health services            - School dropouts                    in the labour market          Poverty
                      social services       - Lower nutrition standard         - Lower quality education            (involuntary)                -Changes in
                      -Lack of social       - Problems with family                                                  - Increasing number of        Income
                      safety net            planing                                                                 street children               Distribution

                                                                   Lower quality of human
                                                                     capital in the future

This section presents the impact of the crisis on labour markets in Indonesia using the
preliminary data available up to February 1999. The major source of data used is the
Sakernas (National Labour Force Survey), which is collected annually by the Central Board
of Statistics. The data used in this section was taken in August 1998, which already captured
the impact of the crisis. The numbers surveyed for the Sakernas 1998 data were about 49,000
households, or around 160,000 individuals. As the impact of the crisis on the labour market
would also depend on the condition of the labour market prior to the crisis, it is important to
discuss labour market conditions during the boom years of the early 1990s.

The labour market before the crisis
The Indonesian labour market from 1990 to 1996 was characterized by a rapid labour market
transformation followed by labour market tightening and increasing real wages. Labour
moved from the informal to formal sector, from rural to urban areas, and from primary sectors
such as agriculture to modern sectors such as construction, manufacturing and services. The
stimulus for labour market tightening was broad-based growth in all sectors, especially the
labour intensive manufacturing sector. The working age population increased by 2.2 per cent
annually from 136 million in 1990 to 154 million in 1996. About 73 per cent of Indonesian
men, and only 44 per cent of women, were in the labour force. About 30 per cent of women
were engaged in household activities, which is counted as not working. A strong pattern of
urbanization can be discerned - the urban working age population grew by 5.8 per cent
annually from 1986 to 1996, compared with a corresponding 0.7 per cent growth in rural
areas. During the boom period of 1990-96, the labour force in urban areas increased by 7.5
per cent annually, compared to 0.8 per cent in rural areas.

Rapid labour market transformation

Labour market transformation in Indonesia in the early 1990s followed the common pattern of
many developing countries, from a labour surplus to a labour scarce economy following
industrial development. This transition marks the ‘turning point’ in the labour market where
wages start to rise faster with the increase in employment as a result of labour market
tightening. This framework is also useful in analysing the bargaining power of labour, the
patterns of income distribution, and the potential changes in industrial structure associated
with economic development. 5

Indonesia’s labour market was undergoing structural transformation and was just entering the
‘turning point’ phase in the early 1990s. The first sign was the rapid decline in employment
growth in agriculture and the remarkable movement of labour from the agriculture to non-
agricultural sectors. The share of agriculture in employment decreased from 56 per cent in
1990 to 44 per cent in 1996. Between 1990-1996, agricultural employment declined at an
annual rate of almost 2 per cent. In Java, employment growth in agriculture declined to -3.4
per cent from 1990 to 1996, whereas outside Java, it declined to near zero in the 1990s. At
the same time, urban employment grew by 7.6 per cent annually, whereas rural employment
grew by only 0.5 per cent.

The second sign was the rapid growth of value added and labour productivity in industries,
especially in manufacturing. The growth of value added in manufacturing averaged more

 See Manning (1995). Some of the features of this transition include a shift from less to more capital
and skill intensive industries, and at the same time a narrowing of the difference in wages between
skilled and unskilled labour.

than 10 per cent per annum from 1971 to 1990. Labour productivity also rose steeply, on
average over 5 per cent per annum outside oil and mining.6 The third sign of transition was
the growing importance of the construction and services sectors. Employment in the
construction sector grew by more than 10 per cent a year, and by 9.3 per cent in the transport,
storage and communication sectors.

Another factor that supported the labour market transformation was the formalization of the
labour market – characterized by the increasing size of the formal sector during the boom
years of the 1990s. Formal sector employment increased from 28 per cent of total
employment in 1990 to 35.2 per cent in 1996, while the share of informal sector employment
decreased from 71.4 per cent in 1990 to 65 per cent in 1996. The process of formalization of
the labour market was more rapid in the urban areas, in line with the trend towards greater
urbanization. The concomitant decline in the informal sector was partly due to the decrease in
agricultural employment. While the importance of the informal sector continued to diminish
as a result of economic development, its size remained large.

Compared with other rapidly growing Asian countries, the turning point of Indonesia’s labour
market was late. One important reason was Indonesia’s greater dependence on natural
resources, especially oil, in the early stages of development. Because of the lack of natural
resources, many other East Asian economies such as South Korea, Malaysia and Taiwan had
adopted outward-looking and export-oriented development strategies since the beginning of
industrialization. Their early economic development was based on cost advantages of labour-
intensive industries, while Indonesia relied on the export of oil and gas. It took the recession
and the oil price drop in the mid-1980s for Indonesia to change its economic development
strategy and to rely more on labour intensive manufacturing sectors for exports.

As part of the labour market transformation, workers experienced a sustained increase in
employment during the economic boom period in the first half of 1990s.7 The usual labour
market indicators such as total employment and unemployment growth, however, do not
provide useful information in Indonesia’s case. The growth of total employment would be
roughly equal to the growth of the working age population because people were considered
working even if they only worked for one hour a week. A better indicator would be the
number of formal employees and employers, which captures the size of formal employment. 8
From 1990 to 1996, the numbers employed in the formal sector grew at around 5.7 per cent
annually, larger than the annual growth of the labour force at 2.5 per cent.

The decreasing number of family workers recorded in the national labour force surveys also
suggests growth in the size of the formal sector. From 1990 to 1996, the number of family
workers decreased annually by 3.8 per cent for females, and 6.8 per cent for males.
Significantly, there was also an increase in female self-employment.

The unemployment rate was low in Indonesia because people in general could not afford to be
unemployed. Following international standards, unemployment was defined as working for
less than an hour a week and at the same time looking for work. In the absence of any social
security scheme, people had to work more than an hour a week to survive, and therefore, the
unemployment rate was generally low, at around 5 per cent or less. In 1996, for example,
open unemployment was almost 5 per cent, and was dominated by unemployed young people.
The unemployment rate was highest among the 15-24 age group (around 13 per cent), which

  See Manning (1995).
  The impact of rapid economic growth on the labour market is discussed in Agrawal (1996).
  In this paper, formal employment is defined narrowly as an employee or an employer with permanent
employees. The informal sector includes the self-employed without any employees, self-employed
with family or part-time workers, and family workers.

includes the age group when secondary school graduates enter the job market.
Unemployment among those who were 30 years or older was generally not significant in
Indonesia. Another characteristic of unemployment in Indonesia was urban unemployment,
particularly among the youth. While urban areas accounted for only 30 per cent of total
employment, they accounted for almost 60 per cent of total unemployment.9

However, in general, unemployment was not considered a serious problem in Indonesia.10
First, those unemployed individuals were mainly young people who entered the job market
for the first time. Second, the duration of unemployment was usually short – an indication of
a relatively flexible labour market in Indonesia.

Although unemployment might not have been significant, under-employment, defined in
Indonesia as working for less than 35 hours a week, was severe. Unlike for unemployment,
there was no apparent relation between economic growth and underemployment – while high
economic growth between 1990-96 significantly reduced unemployment, especially in the
formal sector, the underemployment rate was always high. In 1996, at the onset of the crisis,
35 per cent of workers in Indonesia were underemployed. Underemployment in the rural
areas was much greater than in the urban areas. Around 45 per cent of workers in the rural
areas were underemployed compared with around 21 per cent in the urban areas. Such higher
underemployment in rural areas was primarily due to the seasonal nature of agricultural work.

Wage growth
 Manning’s (1998) analysis of the Indonesian labour market revealed that till the 1990s, the
growth of real wages diverged across sectors. By the early 1990s, there was a more uniform
growth of real wages across sectors. Real wages grew in all sectors in the economy,
particularly in the booming non-oil sectors and tradable sectors such as manufacturing, and
other labour-intensive industries. For example, the real wage in the manufacturing sector
increased by 33 per cent from 1990 to 1996, or at an annual growth rate of around 5 per cent.

As a result of labour market tightening across all sectors, real wages in the agricultural sector
also grew at about the same rate. The wage in the agricultural, forestry, and hunting and
forestry sectors increased by 52 per cent from 1990 to 1996, while that of the community,
social and personal services sectors increased by more than 70 per cent.

An exception was the real wages of civil servants, which after growing at more than 10 per
cent per annum in the 1970s, was relatively stagnant in the 1980s and throughout the 1990s.
The growth of civil service salaries was lower than that of the private sector for the last two
decades. This mirrored the decreasing role of the government in the economy following
deregulation and economic liberalization in the late 1980s and 1990s.

Another notable feature was the growth in real wages for women. In agriculture,
manufacturing, services and even construction, the monthly wages for women grew at a faster
rate compared with that of men. In the manufacturing sector, for example, the real wages of
female workers in 1996 was 62 per cent higher than in 1990. In general, the rise in wages
could be attributed to the sharp increase in minimum wages (which tripled between 1990 and
1996), and the labour market tightening of the late 1990s. For women in particular, the high
growth in wages was due to the increasing participation of women in the formal sector and
longer working hours.

    See Manning and Jayasuriya. (1996).
     See Agrawal (1996)

The Labour market during the crisis: Labour force, employment and unemployment
Labour force
One of the perceptible changes during the crisis was that it increased the size of the labour
force in rural areas. The growth of the labour force in rural areas increased from less than one
per cent a year between 1990-1996 to 2.9 per cent in 1997-1998. At the same time, the
annual growth of the population not in the labour force decreased from 0.2 per cent during
1990-1996 to -3.7 per cent in 1997-1998. Of those not in the labour force, the number of
people attending school in rural areas increased significantly, from an annual growth of
around zero before the crisis to 3.4 per cent from 1997-98. In contrast, the growth of the
labour force in the urban areas decreased from over 7 per cent a year between 1990 to 1996 to
4.7 per cent during the crisis in 1997-98. The growth in labour force participation in rural
areas, combined with the increasing number of people in school in rural areas, suggests that
there was a movement of people from urban back to rural areas during the crisis. This then
suggests that the crisis was basically an urban crisis.

Another perceived trend was the growth of the female labour force and employment. From
1990 to 1996, the number of females in the labour force grew at around 2.3 per cent annually,
and during the crisis in 1997-98, it grew by 4.8 per cent. The growth in female employment
also increased from 1.8 per cent annually before the crisis to 4.2 per cent during 1997-98. On
the other hand, the size of the male labour force remained relatively constant. The increase in
the labour force participation rates for women could be interpreted as the result of an increase
in employment opportunities for women. However, a more likely explanation is that the
crisis forced women to involuntarily participate in the labour market to provide additional
income for the household. The increase in the number of women attending school and the
decrease in the number of those doing the housekeeping reinforces the notion that the crisis
created a larger space for women in the labour market.

  Figure 2.1 Annual Growth of the Labour               Source: Sakernas data, various years
     Force by Urban-Rural, 1990-1998
  4.00                                     1990-1996
  3.00                                     1997-1998
          Rural     Urban      All

Employment reduction

As the Indonesian labour market is dominated by the informal sector, it is very difficult to
precisely measure the extent of the reduction in employment. Measuring employment
reduction in the informal sector is almost impossible because self-employed, temporary and
family workers dominate the sector. Even if only the formal sector, which accounted for
around 35 per cent of total employment in 1996 is considered, there are no reliable data to
measure the reduction in employment or layoffs during the crisis. Firms are not strictly
required to report the number of lay-offs to the Department of Manpower. Therefore, the
number of layoffs recorded in the Department of Manpower – around 151,000 workers from
July 1997 to August 1998 – was clearly an underestimate of the actual figure. Nevertheless, it
is evident that the number of reported layoffs had significantly increased since the crisis


The flexible nature of the labour market, and the high labour turnover rates, even in the
formal sector also complicated procedures to measure employment reduction. Furthermore,
in some industries, especially in the export-oriented, labour-intensive manufacturing
industries, the number of workers tended to be cyclical in nature. For example, it would not
be considered as a layoff when firms did not re-employ workers who voluntarily quit and
went home during the low production period. Workers could also be temporarily laid off, and
such cases were not considered as layoffs.

Before the Sakernas 1998 data were available, Feridhanusetyawan (1998) used employment
elasticity (even though the assumption of constant employment elasticity before and after the
crisis seemed to be too strong) to predict employment reduction. He estimated that a
reduction in economic growth by 13 per cent in 1998 would lead to a reduction in total
employment by around 3 million, and a reduction in formal employment by around 4 million.
This implicitly means that around 1 million workers who lost their jobs in the formal sector
went to the informal sector.11 Using a similar methodology, the ILO predicted that around 5.4
million workers lost their jobs in 1998.12

Another estimate used for employment reduction or layoffs is the decrease in the formal
sector wage employment. The results of the Sakernas 1998 data showed that formal
employment decreased by 4.7 per cent from 1997 to 1998, while it grew by 5.5 per cent
annually during 1990-96 before the crisis. The numbers employed in the formal sector – this
includes both employees and employers – decreased from 31.7 million in 1997 to 30.3 million
in 1998, suggesting that around 1.4 million people left the formal sector during the crisis. At
the same time, informal sector employment increased from 53.7 million in 1997 to 57.3
million in 1998, suggesting that the informal sector absorbed an additional 3.6 million people
during the crisis.

The number of layoffs can also be estimated from the decrease in the number of employees in
the formal sector from 30.3 million in 1997 to 28.8 million in 1998, which could mean around
1.5 million layoffs. A gender disaggregated profile of lay-offs revealed that most of the 1.5
million people who left the formal sector were male workers. The greater impact on males
would be expected, in view of the sectoral contractions in output. The sectors which were
most affected by the crisis, such as the construction sector were primarily male dominated.
Meanwhile, the sectors that survived well during the crisis were mainly female-dominated,
labour-intensive, export-oriented sectors such as the textile, garment and footwear industries.

Employment reduction also occurred mostly in the private sector. The public sector tends to
be more rigid, and adjustments were mainly in the form of wage reductions.

Employment reduction or the level of lay-offs clearly showed a negative trend during the
crisis. At the same time, unemployment also increased as a result of the economic crisis, but
less than initially predicted. The data from Sakernas showed that the total unemployment rate
increased from 4.7 per cent in 1997 to 5.5 per cent in 1998.13 The preliminary result of the
Susenas data was similar – unemployment increased from 5.0 per cent in 1997 to 6.8 per cent

   See Feridhanusetyawan (1998).
   See ILO (1998).
   Sakernas (Survey Angkatan Kerja Nasional – National Labour Force Survey) is a major source of
data about the labour market in Indonesia used in this study. The survey is conducted annually in
August, and the result is usually published in February of the following year. The data used here is for
August 1998, which already captures some of the impact on the crisis.

in 1998. 14 This was much less than earlier government estimates of about 15-20 million
unemployed persons. 15 Later, more reliable official estimates of unemployment in 1998 by the
Ministry of Manpower and Bappenas (National Planing Agency) placed the level of
unemployment at around 14.8 per cent (13.7 million) and 13.6 per cent (12.4 million)
respectively. The special task force of the ILO 16 estimated that the open unemployment rate
in 1998 would be around 10 per cent, or 9.3 million people.

The lower than predicted unemployment figures cannot be taken as an indication that there
were no major lay-offs in the labour market. On the contrary, the lower than estimated
increase in the unemployment rate reflects the rapid absorption of displaced workers from the
shrinking sectors into other industries or the informal sector. Adjustments in the labour
market occurred in terms of adjustments in real wages to the changing supply and demand
situation. Thus, the smaller than expected unemployment figures demonstrate the flexibility
of the Indonesian labour market. In other words, the price of this lower unemployment rate
was a sharp correction through reductions in real wages.

Moreover, another reason for the lower unemployment rate is related to the definition of
unemployment.        Comparable to international standard practice, in Indonesia, open
unemployment is defined as “working for less than one hour a week and at the same time
looking for a job”. Within the prevailing conditions of shrinking real income, stagflation and
no unemployment insurance and social security at the time of the crisis, unemployment was
not a viable option for most people. Hence, even though there were people who were
working for as little as two hours a week, they would not be counted as unemployed. As this
may have included a significant proportion of people, it is not surprising therefore that the
unemployment rate was less than expected.

However, while the aggregate unemployment rate in 1998 was only around 5-6 per cent, the
unemployment rate in the urban areas was three times greater than in the rural areas. The
Sakernas data revealed that the unemployment rate in urban areas increased from 8.0 per cent
in 1997 to 9.3 per cent in 1998, compared to the increase from 2.8 to 3.3 per cent in rural
areas (Figure 2.2).

   Susenas (Survey Sosial Economi Nasional – National Social Economic Survey) is another source of
data that covers a wider range of information such as expenditure. The survey is conducted annually in
February, but the result is usually available for the public much later – sometimes more than one year
later. The sample size of Susenas is four times larger than Sakernas and they are generally not
comparable. The results from the Susenas data in this paper are taken from Sigit (1999).
   From February – May 1998, the Minister of Manpower, Representatives of Labour Unions, and other
observers announced large and overestimated unemployment figures, ranging from 13 to 20 million
people unemployed. They might have used these large numbers to show the seriousness of the
employment problem in Indonesia. See for example, Jakarta Post, April 14, 1998, Suara Karya, May
14, 1998.
   See ILO (1998).

                                   Figure 2.3. Unemployment as a Percentage of
                                             the Labour Force, 1997-1998
                            5.00                                                           1997
                            4.00                                                           1998

                                     Female          Male     Rural    Urban    All

Source: Sakernas data, various years

The Sakernas data also indicated that male workers were more severely affected by the crisis.
Male unemployment increased from 4.1 per cent in 1997 to 5.0 per cent in 1998 while that of
females increased from 5.6 per cent to 6.1 per cent for the same period. As mentioned earlier,
this is because the sectors that suffered the most during the crisis were predominantly male,
while the labour intensive and export-oriented industries that survived the crisis employed
more female workers.

The preliminary results of the Susenas data showed that the majority of the unemployed
labour force consisted of those who had just recently entered the job market. Around 79 per
cent of females and 68 per cent of males who were unemployed had never worked before.
There was no significant rural-urban gap, and this data confirmed that young people
dominated the composition of unemployment in Indonesia. In terms of education, the
preliminary Susenas data suggested that the increase in unemployment was among those with
higher education. The percentage of unemployed people with high school education or above
increased from 1997 to 1998. These results support the notion that only those who had higher
incomes could survive and afford to be unemployed and their unemployment spell could be

There appeared to be two opposing effects of the crisis on underemployment. After the crisis,
the reduction in employment opportunities may have reduced the number of working hours
and increased underemployment. Alternatively, a sharp reduction in real wages and income
also forced people to work more hours, and therefore, could have reduced the number of

                  Figure 2.4. Underemployment in Percentage, 1986-1998
                  15                                                           Less than
                  10                                                           15 hrs
                                                                               Less than
                                                                               35 hrs
                    1986      1990            1996          1997      1998

Source: Sakernas data, various years

The preliminary results from the Sakernas data suggested that underemployment increased in
both urban and rural areas, but the increase was larger in the urban areas which reflects the
severity of the crisis in urban areas. Furthermore, formal employment was much greater in
the urban areas, and the decrease in formal employment forced those who lost their jobs to
work less than 35 hours a week in the informal sector. In relation to gender, there was an
increase in the level of underemployment for both male and female workers.


Earnings and productivity
Accurate data on income is generally hard to obtain in developing countries, including
Indonesia. However, the Sakernas reported the total monthly earnings of wage or formal
employment, which at least presents a picture of monthly wages in the formal sector before
and after the crisis. The data on the monthly earnings of those working in the informal sector
were almost impossible to obtain since informal sector employment was dominated by those
in self-employment and family workers. Nevertheless, it could be inferred that with the shift
of workers into the informal sector, there would certainly have been downward pressure on
earnings in the informal sector.

Nominal and real wages generally moved along the same pattern for the last ten years before
the crisis, with inflation at around 8 to 10 per cent as the difference. However, during the
crisis, with more than 77 per cent inflation in 1998, there was a drastic decline in real wages.
The 1998 Sakernas data showed that nominal monthly wages, measured as monthly
earnings17 in the formal sector, increased by 16 per cent from 1996 to 1997, and by 17 per
cent from 1997 to 1998. Having corrected for inflation, real wages still increased by around 4
per cent in 1997, but then sharply decreased by 34 per cent in 1998. The new level of real
wages in 1998 was comparable to the level in 1989/90, negating the gains in real wages in the
previous seven years. If it is compared with the changes in personal consumption in the GDP,
which declined only by around 3 per cent, the data suggests that people maintained their
consumption level by consuming their savings.

The largest decline in real wages, as given in Table 3.1. - by around 35 per cent - from 1997
to 1998 occurred in public utilities, manufacturing and construction. These sharp reductions
were expected, given the shrinking employment in those three sectors during the crisis.
Unlike employment, which varied between sectors, the decline in real wages appeared to be
more evenly distributed. The smallest decline in real wages was in the mining sector, even
though employment contraction in the mining sector was sharp in 1998. The declining real
wages and employment contraction in some sectors, especially property, financial, banking
and insurance, should continue for some time, due to the slow progress of bank restructuring
in Indonesia.

There has not been any real indication of the difference in the real wage decline between the
urban and rural areas, even though the earnings data in the formal sector from the Sakernas
suggested that the crisis was mainly an urban phenomenon. A quick estimate by using similar
price deflators, indicated that the decline in real wages in urban areas from 1997 to 1998 was
around 36 per cent compared with around 32 per cent in rural areas. The data also showed that
the decline in real wages for male workers in urban areas was larger than that for females.
But, other data suggested that the decline in real wages in rural areas, represented by real
wages in agriculture, was around 40 per cent from 1997 to 1998. Similar trends were
observed in the outer islands, such as Sulawesi, Nusa Tenggara and Sumatra.

     Monthly earning consists of monthly wage plus other benefits.

Classified by the level of worker’s education, those with higher education seemed to suffer a
larger real wage decline compared with those with less education, even though the difference
was not significant. Those who had university and diploma education suffered the largest
decrease in real wages, of around 35.7 per cent. Real wages declined by about 35 per cent for
both workers with secondary education, and workers with primary education. Finally, those
with no education suffered a 33.2 per cent decrease in real wages.

Table 3.1. Growth of Real Wages 1997-98

  Classification                 Growth *)
  Electricity, Gas & Water            -39.1
  Manufacturing                       -37.7
  Construction                        -35.3
  Services                            -33.5
  Trade                               -32.8
  Transport                           -31.1
  Finance & Insurance                 -31.0
  Agriculture                         -26.6
  Mining                              -21.4
  Gender and Urban-Rural
  Urban                               -35.9
     Male                             -36.2
     Female                           -34.4
  Rural                               -31.6
     Male                             -30.4
     Female                           -33.0
  Education level
  Tertiary Education                  -35.7
  Secondary Education                 -35.0
  Elementary Education                -34.5
  No Education                        -33.2
  *) Deflated by 77.6 per cent inflation in 1998.
  Source: Central Board of Statistics, Sakernas, 1997-1998

Labour productivity, measured as real GDP divided by the number of employed, also declined
sharply in 1998 due to economic contraction. During the early period of the crisis in 1997,
labour productivity continued to increase by about 3 per cent, but then fell sharply by around
16 per cent in 1998. At the same time, real wages increased by around 4.2 per cent in 1997,
and then declined sharply by 34 per cent in 1998. While real wages during the crisis were
comparable to that in 1989/90, the level of labour productivity was similar to that in 1993/94.
In other words, real wages declined faster than labour productivity during the crisis, which
opened the possibility of an undershooting wage.

There were differences in labour productivity and real wage adjustments across sectors. The
data from the Sakernas showed that real wages in the manufacturing sector declined by 37.7
per cent in 1998, compared with 26.6 per cent in the agricultural sector. However, because of
the sharper contraction in manufacturing employment, real labour productivity in this sector
remained relatively constant from 1997 to 1998. In contrast, labour productivity in agriculture
declined by 15.2 per cent. Clearly, the biggest labour market adjustment in 1998 in terms of
employment and real wages took place in the manufacturing sector. Although the Indonesian

economic crisis started as a financial crisis, the adjustment in the manufacturing sector was
generally faster than the financial sector.

                         Figure 3.1 Annual Growth of Nominal Wages, Real
                             Wages and Labor Productivity, 1989-1990


                                                                                  No minal Wage
                                                                                  R eal Wag e
                    -10.00                                                        La bo r P ro d.


                             89-   90-   91-   92-    93- 94-   95-   96-   97-
                             90    91    92    93     94 95     96    97    98

Source: CBS, Sakernas, various years

Table 3.2: The annual growth of real wages and productivity

                       1990 1991 1992 1993 1994 1995 1996 1997 1998
All Sectors
  Nominal Wage           16.2 13.4 14.1 23.8         9.7 13.7 15.7         16.2    17.2
  Real Wage               5.7    3.1   8.6 12.4      0.0     4.4     8.5    4.2 -34.1
  Labour Productivity     2.3    4.7   5.9 14.0      1.7     6.6     7.3    3.2 -16.6
  Nominal Wage           13.5 23.5     2.7 29.2 17.9 12.3 13.8             11.5    30.7
  Real Wage               3.2 12.4 -2.3 17.3         7.5     3.1     6.8   -0.1 -26.6
  Labour Productivity -6.4 -1.2        6.0 10.7      7.8     8.2     6.0    7.2 -15.2
  Nominal Wage           29.5    4.5   8.4 34.8      4.1 15.2 17.7         18.4    11.0
  Real Wage              17.8 -4.9     3.2 22.4 -5.0         5.8 10.3       6.1 -37.7
  Labour Productivity 10.1       7.8   8.5 11.0 -10.6 12.9 16.0             1.4      0.1
Source: Calculated based on data from Central Bureau of Statistics, Sakernas, various years.

There were several contributing factors to the specific nature of labour market adjustment in
Indonesia. The first was the flexible nature of the labour market itself, where the labour
turnover rate was generally high. Combined with the labour surplus nature of the economy,
weak labour unions and the low reservation wage, the flexible nature of the labour market led
to flexible real wages. The second factor that also played a major role in sharp wage
adjustment was the specific overall macroeconomic condition, where excessive monetary
expansion led to high inflation and sharp depreciation. This specific macroeconomic
condition characterized the Indonesian crisis and in turn contributed to slightly different
adjustments in the labour market compared with that in Thailand or Korea. In those two
countries, inflation was still in single digits and the real wage was relatively stable, which
meant that the adjustment occurred mainly in the form of rising unemployment rates.

Labour market transformation
With the economic crisis, the process of labour market transformation slowed down, and it

                         Figure 3.2. Annual G row th of Em ploym ent By Sector,
                          20. 13.
                          0                       10.         9.
                               0  6.   5.                                     6.
                          10.                3.   7     6.5 3                         4.
                          0       6    8                  -      0.           3       4
                           0.                             0.8    7
                           0 -                                                          -

                         -    1.9                                               -       1.4
                         10.             -                                      5.8
                         -               9.8        -
                         20.       -                15.
                         -                    -
                         40.                  36.

could even be reversed as economic performance continued to deteriorate in 1999.

First, labour no longer moved from the agricultural sector to the manufacturing and services
sectors as formal employment in the manufacturing and services sectors decreased
significantly. Since the start of the crisis, there was a massive reverse movement of labour
from the non-tradable sectors, especially manufacturing, construction and the inefficient
financial sectors to the agricultural sector. Fiscal budgetary constraints reduced government
spending on infrastructure such as roads and water irrigation, and this prevented further
expansion of labour demand in the construction and utilities sectors. The depreciation of the
rupiah improved the terms of trade in agricultural commodities, so that the booming
agricultural sector could absorb displaced labour from other sectors.

The 1998 ILO report predicted this sectoral reallocation of workers, but did not provide
precise estimates. The report estimated that around 20 per cent of workers, both in formal and
informal activities, would be displaced in the manufacturing, trade, transport and service
sectors. Worker displacement in the construction and financial sectors was estimated to be
around 30 per cent.

The result from the Sakernas data (figure 3.2) showed that during the crisis in 1998, there was
a dramatic decline in employment in the utility (electricity, gas and water), mining,
construction and manufacturing sectors by about 37 per cent, 23 per cent, 16 per cent and 10
per cent respectively. The collapse of the financial sector during the crisis also led to a
reduction in employment by 5.8 per cent in 1997-98. During the economic slow down,
agriculture turned out to be the saviour of the economy with an increase in labour absorption
(or employment) by 13 per cent during the crisis period. After decreasing from 56 per cent in
1990 to 41 per cent in 1997, the share of employment in the agricultural sector increased to 45
per cent in 1998.

Also, an interesting phenomenon following the crisis was the sharp increase in agricultural
activities in urban areas – a leap from the previous annual growth (during 1990-1996) of 2.2
per cent to 45 per cent. Part of the explanation lies in the utilization of unused property lands
in the urban areas – also known as ‘lahan tidur’ or ‘sleeping land’ – for agricultural purposes
as a result of the crisis.

Source: CBS, Sakernas data, various years.

However, this process of reverse movement from manufacturing to agriculture cannot be
expected to continue in the longer run, as agricultural liberalization will lead to more capital-
intensive and efficient agriculture. Hence, the sector cannot always function as a labour-
absorbing sector.

A second reason for the slowdown and possible reversal of labour market transformation is
that labourers moved to rural areas because of the crisis, and hence, the process of
urbanization slowed down. The crisis reduced the growth of urban employment from around
5.2 per cent per annum during the boom years in 1990-96 to around 3.1 per cent from 1997-
98. On the other hand, from an annual growth of 0.6 per cent before the crisis, rural
employment increased to 2.3 per cent after the crisis. While urbanization will not be a
serious problem in the near future, the problems related to urban poverty are expected to
continue. Those unskilled construction and manufacturing workers in the urban areas who
lost their jobs during the crisis and who also lost their ties to their rural origins were the group
who were badly hurt by the crisis, and contributed to the problems of urban poverty.

Third, there was a movement of labour from the formal sector back to the informal sector,
especially in the urban areas. The share of the informal sector in general increased by 6.9 per
cent (from 63.3 to 65.4 per cent) from 1997 to 1998. This increase was much larger than the
annual growth of the informal sector at around half a per cent in the decade prior to the crisis.
At the same time, the growth of formal sector employment declined by -4.4 per cent.

The informalization process was strong both in the urban and rural areas, but was particularly
notable in the urban areas. Prior to the crisis, during 1990-1996, people were leaving the
informal sector at the rate of 2.9 per cent in urban areas, and 1 per cent in rural areas
respectively. After the crisis, there was a positive movement into the informal sector, which
registered a substantial growth of 10.3 per cent in urban areas, and a slightly lower growth of
5.8 per cent in rural areas. Meanwhile, there was a corresponding contraction in formal sector
employment. This decline was more marked in rural areas, where formal sector employment
decreased by -7.7 per cent, compared to - 2.1 per cent in urban areas.
                     Figure 3.3. Share of Employment By Status within the Informal
                                            Sector, 1986-1998



                                                                                     with no employee
                                                                                     with temp/family
                    25                                                               workers
                                                                                     Family workers

                                  1990         1996        1997       1998
Source: CBS, Sakernas data, various years

A review of the employment status within the informal sector indicates that the crisis reduced
the proportion of the self-employed with no employees, and increased the share of family
workers. The number of self-employed with no employees still increased from 19.7 million
to 20.5 million, but the number of self-employed with family workers increased from 18
million people in 1997 to 19.7 million in 1998. Increased employment in the informal sector
was expected, but the fact that the number of family workers increased significantly during
the crisis suggests that people looked for jobs within their family unit as part of the coping
mechanism to survive the crisis.

Another noteworthy trend was the increase in the number of female family workers during the
crisis. Before the crisis, from 1986 to 1996, the number of female family workers decreased
by 1.8 per cent annually. This reflected greater opportunities for women working in the labour
market outside the family. However, from 1997 to 1998, the number of female family
workers increased by 6.5 per cent. Family workers accounted for 36 per cent of total female
workers in 1998. There was also a significant growth (by around 10. 8 per cent compared to a
negative growth of 3.2 per cent before the crisis) of male family workers, but they accounted
for only around 9 per cent of total male workers.

Combined with the sharp reduction in real wages – up to 30 per cent in general – all these
signs show the rapid reversal of labour market transformation in Indonesia as a result of the

The reversal of the labour market transformation means that Indonesia is moving back to a
labour surplus economy at least for some time in the near future. The structure of the
Indonesian labour market has moved back to a situation similar to the late 1980s. The new
level of real wages in 1998, which was similar to the level of the late 1980s, and the new level
of labour productivity that was comparable to that in the early 1990s reflected this reversal

However, the slower process of labour market tightening increased the competitiveness of
Indonesian manufactures in the international market, at least in the short run, because of two
forces – more favourable terms of trade (in relation to wage standards) due to currency
depreciation, and adjustments in real terms when labour productivity growth became higher
than real wage growth.

As mentioned in the previous section, the process of labour market transformation and turning
point has been a longer process compared with that in other Asian countries. The crisis not
only further delayed labour market tightening but also lengthened the process of catching up
with other neighbouring countries. For example, the growth in real wages in manufacturing
in Thailand increased by around 70 per cent in the last 10 years, while that of Indonesia
increased by a mere 30 per cent. In Thailand, inflation has been under control, so that the
sharp decrease in real wages could be prevented. But with the economic crisis in Indonesia,
the one-decade increase in real wages was diminished within one year, between 1997-98.
Another factor that potentially led to a slower transformation process was the lower level of
education of the Indonesian worker compared with workers in other Asian countries. The
lack of education hindered the process of industrial transformation because people were not
flexible enough to change jobs to a more modern sector where education and skills were


International Labour migration
Despite the lack of reliable data on recent trends, it appeared that the crisis increased the
number of Indonesian workers working overseas. According to some newspaper reports, the
number of Indonesian workers working overseas (or the TKI = Tenaga Kerja Indonesia or
Indonesian Labourers), and their foreign exchange earnings increased significantly from 1997
to 1998, but the precise number seemed to be conflicting. The government was promoting the
programmes of sending Indonesian workers overseas, and in 1998, the number of the TKI was
slightly above the government’s target at around 1.25 million. Malaysia received the largest
number of TKI, with 423,625 workers up to May 1998. Saudi Arabia was the second most
popular destination with 417,665 TKI up to June 1998. However, the type of TKI in these
two countries was markedly different. In Malaysia, around 70 per cent of the TKI worked in
the formal sector, while in Saudi Arabia, 70 per cent of the workers were women, most only
had primary education, and worked in the informal sector. The majority of Indonesian
workers working in Saudi Arabia were domestic helpers. The next popular destinations were
Singapore and Hong Kong, but most of the TKI that worked in these two countries were the
better educated. On average, they had a high-school education.18

While data on illegal migrants is not easily available, it is generally understood that most of
the illegal workers from Indonesia chose to go to Malaysia because of the distance and the
similarities in language and religion. The Ministry of Manpower in Malaysia estimated that
until November 1998, there were 72,191 illegal workers from Indonesia working in
Malaysia.19 It was also reported that in 1997, Malaysia deported 38,153 illegal Indonesian
workers, and until October 1998, Malaysia had again deported 52,000 workers from
Indonesia.20 This shows that there was an increase in the number of Indonesians crossing the
border and working illegally in Malaysia following the crisis. It was estimated that there were
around 1,700 illegal Indonesian workers entering Malaysia from Sarawak (in Kalimantan
Island) every month. 21 It was estimated that monthly, around 6,000 illegal workers crossed the
Malacca Straits border by boats (Pongpong).22

Table 3.3. The number of foreign workers in Indonesia by main occupation

  Classification           1996              1997               1998
  Executives             12,663             8,762              9,497
  Professionals          11,163            12,969              7,206
  Supervisors             8,281             5,409              3,864
  Technician/Operators   16,551            10,052             12,864
  Total                  48,658            37,192             33,431
 Source: Department of Manpower website (, taken in July

During the boom period of the early 1990s, the number of foreign workers in Indonesia
increased significantly. Most of them were professionals and executives, hired mainly
because of the scarcity of skilled labour in the domestic market. The depreciation of the

   Data are from Depnaker, quoted from Suara Pembaruan, “Permintaan TKI meningkat”, July 11,
   Bisnis Indonesia, “Calo Puas Memeras, Pekerja Memelas”, November 6, 1998
   Republika, “Malaysia Mendeportasi 10.000 TKI Ilegal”, March 21 1998.
   Kompas, “1700 per bulan, TKI Masuk Sarawak”, April 23, 1998
   The Jakarta Post, 'Steps to protect workers 'slow compared to promotion'", December 18, 1998

rupiah clearly reduced the ability of domestic companies to employ foreign workers. The data
from the Department of Manpower indicated that during the boom years in 1996, the number
of foreign workers in Indonesia was around 48,700. In 1997, the year when the crisis started,
the number decreased to about 37, 200 and in 1998, it decreased further to around 33,400.
From 1997 to 1998, the number of foreign professionals declined significantly from around
13,000 to 7,200 workers.


Regional distribution
In view of the diversity of regions in Indonesia, an assessment of the regional impact of the
crisis is a complex task, both in terms of urban-rural, and also Java and off-Java

With regard to the urban-rural differences, as discussed earlier, most studies point towards an
urban bias of the crisis. This is further highlighted by the decrease in average spending in
urban areas by 34 per cent compared to 13 per cent in rural areas in 1997 to 1998, which
coincides with the decline in urban employment during this period.23

Meanwhile, a report based on the Kecamatan Rapid Poverty Assessment survey revealed that
both rural and urban areas in Java were the most severely affected by the crisis. The survey
measured the “coping index” which was based on the indicators of the degree to which people
were selling assets to meet their basic needs, reducing their participation and contribution to
social activities, and other indicators of the use of “coping” mechanisms. The survey placed
all areas in Java within the 20 hardest hit areas, regardless of urban or rural status.

On a regional basis, according to the Kecamatan Survey data, 24 the impact of the crisis on off-
Java islands was heterogeneous, and did not seem to correlate with the level of income or the
poverty level before the crisis. Some areas that were affected by the drought in 1997 actually
escaped from the drought and experienced a booming economy from export crop earnings due
to currency depreciation. Other areas such as East Timor, and East and West Nusa Tenggara
experienced negative effects from the crisis, but it was difficult to ascertain whether the
problems were related to the crisis or to the long drought.

However, an overwhelming common feature of all the studies was that there was no
correlation between the pre and post crisis poverty levels. For instance, although Java had a
low poverty incidence before the crisis, it was the hardest hit during the crisis.

The data on real expenditure also revealed a similar pattern. Those regions with financial,
construction and other modern sectors like Jakarta and West Java were most affected during
the crisis. The average real expenditure per capita, which was well above the provincial
average, decreased by 30 per cent in Jakarta, and 42 per cent in West Java, compared with
only a 10 per cent decrease in North and South Sumatra.

However, the result of the Sakernas 1998 did not significantly show the severity of the impact
in Java compared with areas outside Java. In terms of employment reduction, for example,
the growth in employment varied greatly across provinces, and did not show any significant
differences between Java and off-Java regions. However, this was expected because the
labour market was flexible. Workers might have moved between sectors in the formal sector,
or even to the informal sector.

   The data is from the Indonesian Family Life Survey (IFLS) which is a longitudinal household and
community survey conducted as a collaborative effort between RAND, Lembaga Demografi-UI and
UCLA (University College, Los Angeles). The most recent survey conducted after the crisis is the
IFLS 2+, a detailed resurvey of almost 2,000 households in seven provinces following the economic
crisis (between August and December 1998). For the complete report, see Frankenberg, Thomas, and
Beegle (1999).
   The Kecamatan Rapid Poverty Assessment (RPA) was a subjective, expert respondent survey of
three government officials in each of Indonesia’s 4025 kecamatans carried out by the Central Bureau of
Statistics. The survey selected three respondents with kecamatan-wide responsibilities and asked them
a standard set of questions which tried to measure the different kinds of effects in each of their areas of
responsibility. See Sumarto, Wetterberg, and Pritchett. (1998).

The impact of the crisis on the regional labour market was also complex, considering the
differences in employment structure across provinces and islands in Indonesia. The inter-
related employment indicators, such as unemployment and underemployment, as well as the
formal-informal structure of the labour market, reflected the degree of economic development
in the provinces. There were large differences in the share of the informal sector across
regions in Indonesia. Being the most modern labour market, Java had the smallest proportion
of informal sector employment, at less than 60 per cent. Yet, in some provinces in eastern
Indonesia, for example, informal sector employment accounted for up to 80 per cent of total
employment.       When the proportion of informal sector employment is high, the
underemployment rate is generally high but open unemployment is relatively low. Due to the
different structures of the labour market, the comparison of labour market indicators across
regions would lead to difficulties in interpretation. Nonetheless, it is clear that there was a
larger increase in the informal sector outside Java, compared with that in Java after the crisis.

Table 4.1. Regional employment and wage growth, Selected Regions, 1997-1998

     Region                Employment Real Wage
                            Growth *) Growth **)

Java (Average)                   0.9            -34.9
  DKI Jakarta                   -1.5            -33.4
  West Java                      2.9            -33.3
  Central Java                   2.1            -31.2
  Yogyakarta                    -1.8            -39.0
  East Java                      3.0            -37.4

Sumatra                     5.1          -31.4
Kalimantan                  1.6          -27.7
Sulawesi                    2.7          -24.3
Maluku-Irian Jaya           6.9          -27.4
Bali - Nusa Tenggara       -1.1          -29.6
East Timor                  6.0          -40.8
Source: CBS, Sakernas, 1997-98
*) Total Employment; **) Monthly earning of wage employment divided by regional CPI

In terms of the real wage decline, the data showed that in general, Java experienced larger
wage reductions. Other provinces outside Java that also experienced large wage reductions
included Aceh and North Sumatra.

While the evidence seemed to suggest that Java was badly hurt by the crisis, some reports, for
example by UNSFIR (1999) warned that the eastern provinces that had a high poverty
incidence prior to the crisis appeared to retain this unenviable status in 1998-99. It was clear
at the end of 1998 that despite the emergence of the new urban poor in Java following the
crisis, policies should not overemphasize this new poor and should not be biased against
existing poor areas, especially in eastern Indonesia.25

Distribution of rich and poor
Based on expenditure analysis26, preliminary results of the World Bank Report (Poppele,

   See UNSFIR (1999).
   Precise data on income are generally difficult to obtain, so that measuring the distributional impact
on the rich and poor was generally based on expenditure levels, with all the associated weaknesses
relating to analysis.

Sumarto, Pritchett, 1999) using the IFLS data showed that the average real expenditure
distribution fell more than the median, implying that the distribution of expenditure was more
even after the crisis. The data revealed that the upper middle class was most affected in
relation to a decline in expenditure.

Table 4.2. Changes in per capita expenditure by initial quartile per capita expenditure

          Quartile in        Percentage        Percentage
            1997           change in mean       Change in
        I (low)               120.0                 49.0
        II                     41.9                 11.7
        III                    -1.6                -18.8
        IV                    -54.0                -41.4
        Source: Poppele, Sumarto, and Pritchett (1999)

Classified by education level, the change in expenditure of the household also followed a
similar pattern. The mean of expenditure for those who had no education decreased by 14 per
cent during the crisis, while the mean of expenditure for those with tertiary education fell by
23 per cent. If education level is linked to income, this could be interpreted as a greater
reduction in the expenditure of the rich compared to the poor. Moreover, this corresponded
with the larger real wage decline of those with higher education compared with those with
less education.

However, changes in expenditure cannot be taken as an adequate comparison of declines in
the quality of life. For instance, even though the expenditure of the rich decreased more than
that of the poor, the impact may have been vastly different. For the rich, a sharp decline in
expenditure, could mean switching from having overseas to domestic vacations, while for the
poor, a small decline in expenditure could mean reducing their consumption from eating three
times to twice a day.

Changing patterns of expenditure and income
The changing pattern of income and expenditure at the household level is at the heart of the
analysis of how the crisis affected the lives of people in Indonesia and for understanding
labour market adjustments. Unfortunately, the data on income are generally not available,
even though the data on expenditure has been relatively more reliable and available.

The crisis clearly changed the composition of expenditure as people adjusted their patterns of
consumption in order to maintain their previous levels of well being while facing a sharp
increase in prices, especially food. Food prices increased by 118 per cent in 1998, while
those of clothing and health services increased by 98.7 and 86.1 per cent, respectively.

To understand the impact of the crisis at the household level, a hypothetical framework,
involving the components of the income and expenditure relationship (presented as an
equation) is used below:
                                      Income         =        Expenditure
                                      (+30 %)                       (+30 %)

                Salary + Interest + Profit or Dividend =         Consumption + Saving
                (+17%)          (+13%)                        (more than 30%) (decrease)

(Wage x Hours of work) + Interest +Profit or Dividend =          (Prices x Quantity) +Saving
(less than 17%) (work more)    (+13%)                            (+78%)            (decrease)

The numbers in brackets illustrate the estimated increase or decrease of a certain component
of income or expenditure. Based on various micro data that were available, it is fair to
assume that on average the household nominal expenditure increased by around 30 per cent
from 1997 to 1998. Assuming an identity of income and expenditure, the total household
income in nominal terms should have increased by a similar amount at 30 per cent.

From the income side, the total household income could be broken down into three
components: salary, interest income, and profit or dividend. On average, according to data
from Sakernas, for example, nominal earnings increased by around 17 per cent, which means
that in order to meet the 30 per cent increase in total income, the contribution from the
increase in the interest and dividend income combined should be around 13 per cent. Interest
income in general increased during the crisis because of the high interest rate environment in
the economy.

The salary component could be further broken down into more detailed components, namely
the hourly wages multiplied by hours of work. The increase in wages was generally less than
17 per cent from 1997 to 1998, which meant that the household or the workers would have to
work more hours to meet the 17 per cent increase in salary. For some poor households, the
increase in number of hours of work could mean that more family members, including
children, also participated in the labour market.

From the expenditure side, the increase in total expenditure was partly financed by the
reduction in savings, or even dissaving or selling assets. The consumption component of
expenditure could then be broken down into price multiplied by quantity. The fact that prices
generally increased by almost 77 per cent, while at the same time expenditure increased by
only 30 per cent from 1997 to 1998, suggested that people reduced their quantity and also
quality of consumption.

Table 4.3. Predicted changes in the structure of household real income

  Income     Salary                         Saving        Profit/Dividend     Overall Impact
  Group      Wages          Hours of        Interest                          (Magnitude)
  Poor       Decrease in    Work more       No saving     No dividend         Real Income
             Real Wage                                                        Decline
  Middle     Decrease in   Work more or     Increase      No dividend         Real income
             Real Wage     less depending                                     decline (Small)
                           on interest
  Rich       Wages and salary are small     Increase      Large               Real Income
             component of household                       Decrease            Decline
             income                                                           (Largest)

The hypothetical analysis becomes more complicated when different income groups (Table
4.3) are considered, as the possible adjustment in household income would vary accordingly.
For poor workers, the only source of income would be salary, so that all the required
adjustments would be in the form of working more hours or forcing more members of the
family to work in the labour market. This partly explains why the unemployment rate was not
as high as expected. For the middle income group, the interest income increased during the
crisis so that the total real income decline could be smaller that the reduction in real salary.
The increase in interest income created a buffer to reduce the pain of real income decline. For
the rich, the largest adjustment took place with the collapse of firms, companies, stock
markets and other forms of declining financial assets in their total real income.

Table 4.4. Predicted changes in the structure of household real expenditure
  Income                    Consumption                         Saving            Overall Impact
  Group     Basic Needs           Secondary Needs                                   (Magnitude)
  Poor      Increase in           Decrease;                Dissaving or        Decrease in
            proportion            Substituting for basic   Selling asset       Real Expenditure
                                  needs                                        (Smallest)
  Middle    Increase in           Consuming lower          Saving less or no   Decrease in
            proportion            quality goods, less      saving              Real expenditure
                                  imported goods
  Rich      Increase (smaller)    Large reduction in       Saving less         Large decrease in real
            Or might decrease     luxuries                                     expenditure
            (quality adjustment)                                               (Largest in magnitude)

The possible decomposition of household expenditure is presented in Table 4.4, by separating
consumption into two components: basic items such as food, and secondary items such as
clothing, recreation and others. For poor households, when prices went up by almost 80 per
cent in 1998, and the consumption-based expenditure (prices x quantity) increased by less
than that (approximately around 35 to 50 per cent), there was clearly some adjustments in the
quantity or quality of goods purchased by the household. Since the share of food expenditure
for the poor is generally higher than that of the average consumer, and the increase in food
price was actually 118 per cent, the poor consumer actually faced price increases much higher
than 80 per cent in 1998. With small or even no household savings, all of the increase in
general expenditure would be financed by selling assets, working more hours or substituting
consumption of secondary items to more basic commodities such as food.

For the richer households, interest income and savings affected adjustments. To compensate
for the larger increase in expenditure, compared to that of income, people consumed their
savings, especially those who thought the crisis would only be temporary. From the income
side, the existence of interest income from saving also reduced the chance of working extra
hours to balance the increase in expenditure. In the short run, the richer households had more
leeway in consumption, by selling and consuming their assets as well as reducing their
savings. Nevertheless, this smooth consumption could not be maintained over the longer
term, and with high inflation and economic contraction, they were also forced to reduce
consumption and save more. Some households benefited from the cushioning provided by
foreign denominated assets, which increased sharply in value during the crisis due to the
currency depreciation.

The dynamic adjustment in income for small-scale self-employed individuals in the informal
sector was even more complicated as profits and return to labour input were often mixed. An
increase in price due to inflation would generally increase their revenue, but it did not
necessarily translate into profits due to the increase in the price of purchased inputs. The
increasing number of people working in the informal sector increased competition, which in
turn reduced profit.

This analysis suggests that in order to meet the increasing level of expenditure caused by the
crisis, the wealthy could consume their savings, while the poor had to work longer hours or
even force other family members to participate in the labour market. Therefore, those who
suffered the most from the crisis were the ones with no savings and those who could not
participate in the labour market, especially the disadvantaged groups such as the elderly,
single women with children, people with disabilities and others.

Changes in employment patterns, real wage reductions and labour market trends were related
to the shifts in poverty during the crisis. The estimates of the impact of the crisis on poverty
in Indonesia were beset by controversy. Initially, the ILO27 had predicted that one in every
two Indonesians would live below the poverty line in 1998. The Badan Pusat Statistik
(BPS)28 made a similar large prediction of poverty at around 39 per cent in mid-1998. Given
that the official poverty incidence was only around 11 per cent in 1996,29 these new estimates
at 39-48 per cent showed that that the number of poverty-stricken increased by four times
within two years. These levels were comparable to that in the 1970s, which implied the total
reversal of the development gains of the past 25 years. However, by the end of 1998,
estimates based on different methodologies and the data that were collected during the crisis
showed that the poverty incidence was not as severe as previously predicted.

The initial high predictions by the ILO and the BPS may have been due to the pressure to
estimate the poverty levels without proper data. While the lack of data was clearly a problem,
the major problem in the estimations was related to the assumptions. The ILO, for example,
assumed that the average wage and household income would not change in nominal terms
during the crisis, while various data later showed that the nominal wage increased by around
17 per cent, and household nominal expenditure increased by around 30 per cent. Similarly,
the BPS, based on past income growth, wrongly assumed that those at the lower end of the
income distribution would be hardest hit by the crisis.

While the BPS and ILO applied the ‘doomsday scenario’, another poverty estimate by the
World Bank (calculated by the expenditure data from Susenas 1996, with the measurement
based on the predicted sectoral growth of the economy) indicated a more optimistic figure at
around 14 per cent. By assuming overall economic growth at -12 per cent in 1998, and zero
per cent the year after, it estimated that the poverty figure would be 14 and 14.5 per cent in
1998 and in 1999/2000 (Cameron, 1999). 30 Given that the economic contraction in 1998 was
worse than -12 per cent, this poverty estimate was too optimistic, even though in fact it was
closer to the estimates of poverty at the end of 1998.

   The ILO defined the poverty standard in 1998 as the poverty standard in 1996 x a 16.5 percent
increase in median consumption x an 80 percent inflation ate.
   The poverty standard according to BPS was equal to 2100 calorie per day plus other basic needs,
which was translated to Rp. 52,470 per month per capita in urban areas, and Rp. 41,588 per month per
capita in rural areas.
   The incidence of poverty in Indonesia had declined significantly from around 40 percent in the mid
1970s to around 22 percent in 1984 and just around 11 percent in 1996.
   Cameron (1999).

Table 4.5. The proportion of the population living below the poverty line in Indonesia

                                As a per cent of the population              In million people
           Year                 Urban       Rural       Total       Urban        Rural         Total
1976                             38.8        40.4         40.1       10.0         44.2          54.2
1980                             29.0        28.4         28.6        9.5         32.8          42.3
1987                             20.1        16.1         17.4        9.7         20.3          30.0
1990                             16.8        14.3         15.1        9.4         17.8          27.2
1993                             13.4        13.8         13.7        8.7         17.2          25.9
1996                              9.7        12.3         11.3        7.2         15.3          22.5
BPS Mid-year,
based on declining real
income                          28.8        45.6        39.1        22.6       56.8             79.4
based on declining real
income                          39.3        53.2        48.3        28.1       70.7             98.8
SMERU Report (World
IFLS Data, various deflators 12.0 - 15.8 15.2 – 23.0 13.8 – 19.9 9.1 – 12.0 19.3 – 29.3      28.4 – 41.3
SMERU Report (World
100 Village Data, various                            14.4 – 18.6                             29.4 – 38.0
Susenas data – CPI deflator
1998                            17.2        19.0        18.3        13.1       24.2             37.3
M. Ikhsan*)
Susenas data - Other
methods                         10.6        25.7      20.1**)       7.9        32.1             40.0
1996                            20.3        41.3      33.5**)       15.5       52.6             68.1
BPS, new          poverty
standard                         21.9        25.7        24.2        17.6        31.9           49.5
1998              (officially
Sources: Badan Pusat Statistik, various reports, ILO report (1998), Poppele, Sumarto and Pritchett (1999).
*) Estimates by M.Ikhsan, as reported in UNSFIR(1999)
**) Calculated based on the number of population in urban and rural areas

On 9 July 1999, the BPS announced the new official poverty figures based on the core
Susenas taken in December 1998 - 24.2 per cent of overall poverty incidence, 21.9 per cent in
the urban areas, and 25.7 per cent in the rural areas. It reported that urban poverty increased
more rapidly than rural poverty, showing that urban areas were hardest hit by the crisis. This
official BPS report mentioned that poverty increased from 11 per cent (using the old standard)
to 24 per cent (using the new standard). However, unfortunately, the report did not present the
revised poverty figure in 1996, so that it was actually irrelevant to compare those two

Clearly, the crisis increased poverty incidence in Indonesia although the precise measure
depended on the methodology, price deflators and sources of data that were being used. From
all these various estimates, it would seem fair to conclude that the number of people living
below the poverty line, or the incidence of poverty, increased by around 65 per cent from
1996 to 1998. In addition, the poverty figure for 1998 should be comparable to that in the
early 1990s. The fact that the preliminary data of 1998 GDP showed a decline in personal
consumption expenditure by around 3 per cent, or a level similar to that in 1993, seemed to
support the findings that the poverty figures were similar to the level of the early 1990s.

This generalization, however, should be treated with caution. The measurement of poverty
incidence based on expenditure data has several important implications. First, the
measurement does not capture the case of consumption smoothing, when people spend their
savings or dissaving to maintain their previous level of consumption. Second, the expenditure
data does not capture the changes in the quantity and quality of consumption as a result of
price changes. The different results from various studies on poverty also show the sensitivity
of poverty incidence to the methodology, price deflators, and source of data used in the
survey. In terms of methodology, there are several methods of poverty measurement such as
head count ratio, food poverty lines and others. In terms of price deflators, there could be food
prices, general prices, wholesale prices, or others. In terms of the data, it could be Susenas,
IFLS or others. In addition to those, part of the problem is also purely statistical in nature.
Because of a skewed income or expenditure distribution, where the number of people with
lower income is large, a small change in the poverty line would lead to a large difference in
the number of people living below the poverty line. In other words, the increase in the
poverty line by Rp. 2,000 or US$ 25 cents, for example, would lead to additional millions of
people living below the poverty line.

Furthermore, the aggregate poverty measure masks the complex regional, sectoral and other
distributional dimensions of poverty. Therefore, the more crucial issue related to poverty is
not the precise aggregate measurement of poverty incidence, but an analysis of the difference
in the nature, severity, depth and profile (type of people affected) of poverty before and after
the crisis.

Vulnerable groups

From the previous discussion related to income and expenditure adjustments, it is clear that
those who had no savings and no interest income, or had no capacity to get additional income
through participation in the labour market, were the ones who suffered most from the crisis.
Although reliable data on vulnerable groups was not widely available, children of poor
families or the elderly, who lived near or below the poverty line, and unskilled and single
female labour with children, appeared to be the worst victims. Nonetheless, generalizations
should not be made with regard to the vulnerable groups based solely on gender and age,
especially because of the heterogeneous nature of the social impact. For example, those who
belonged to the vulnerable group before the crisis could be better off after the crisis if in fact
their sources of income depended on export earnings. In other words, a new type of
vulnerable group could emerge as a result of the crisis. This section presents some early
findings of the impact of the crisis on the vulnerable, focusing on women, children and the
The crisis increased both the labour force participation of, and employment for women, but it
is rather difficult to provide a normative judgment on whether this was actually a positive or
negative development for women. More research is needed to better understand the impact of
the crisis on gender dynamics in the society, especially at the micro or household level to
understand the gender impact of the crisis.

The increasing role of women in the labour market during the crisis was clearly the result of
push and pull forces both from the demand and the supply side. From the demand side, the
surviving sector during the crisis turned out to be the labour intensive, export-oriented sectors,
which were considered more ‘feminine’ or female oriented, and the collapsing sectors turned
out to be the male dominated or ‘masculine’ sectors, such as construction, mining and
utilities. From the supply side, the declining real income of households forced women in the
family to enter the labour market to seek additional income for the household, either in the
formal or informal sector.

The sudden increase in the female labour force participation rate and the informal nature of
the work reflected the involuntary nature of this participation. Within the prevailing
household division of labour, women were responsible for managing expenditure and
preparing food. With the simultaneous rise in prices and the decline in family income, there
was clearly a stronger economic pressure for women to maintain the welfare and well being
of the family by switching expenditure and adjusting consumption. When adjustment in
quality and quantity was no longer possible due to the sharp increase in prices, women had to
seek sources of additional income to maintain the expenditure level.

As a result of the general trend that the decrease in male employment was larger than the
decrease in female employment, job creation programmes as part of the social safety net were
tailored more for male workers, such as digging city ditches, building roads and the like. In
other words, the opportunity for women to participate in these public works programmes was

The crisis increased the number of female workers working in the informal sector, which led
to a weakening in the position of women in the labour market. Despite the formalization of
female employment in the last decade, the proportion of female workers in the informal sector
was still larger than that of the males. Because of the part-time and temporary nature of these
informal activities, the proportion of underemployment among women was also larger, and
therefore their income was also smaller. After the crisis, the proportion of women in the
informal sector increased to more than 70 per cent, higher than the overall proportion of the
informal sector at less than 60 per cent. The proportion of underemployed female workers
also increased to more than 50 per cent in 1998, compared with around 40 per cent of the
national average. It is difficult to measure the real income decline in the informal sector, but
higher competition in the informal sector certainly would create a downward pressure on their
wages. The more serious implication, however, was that the increasing number of women in
the informal sector led to a weaker employment position for women. At the moment, there is
clearly no protection for labour, such as minimum wage, labour standards, employment safety
measures, and other important labour protection, in the informal sector.

The immediate impact of the crisis on children was in the form of decreasing enrolment rates
and increasing school dropout rates – especially in secondary school. Parents pulled their
children out of school both to reduce expenses (to cope with the fall in real income after real
wage cuts), and also to participate in the labour market, and provide additional sources of
income. The data in 1998 suggested that the level of enrolment dropped to the level of the
early 1990s. The nation-wide Susenas 1998 did not provide convincing evidence on the
changes in enrolment rates, but the smaller size IFLS data clearly showed a significant
increase in dropout rates among children. The increase was more pronounced for males than
females, for urban residents compared to rural residents, and for children from the poorest
households.31 The dropout rates of children between 13 and 19 years of age (secondary

     Frankenberg, Thomas and Beegle (1999).

school) in urban communities, for example, increased from 11.1 per cent to 17.5 per cent
from 1997 to 1998, compared to an increase from 13.5 per cent to 16.8 per cent in rural areas.
From the IFLS data, it was estimated that the gap in enrolment between the top and bottom of
per capita expenditure quartiles had widened since the crisis, with children from the poorest
households about ten times more likely to be out of school than their counterparts from rich
households. These data confirmed the government’s previous estimation that 2.5 million
children were under threat of being taken out from school, and there would be an increase in
dropout rates.32

Linked to these changes in school enrolment and dropouts were trends in children’s
participation in the labour market. Since 1998, the legal working age was increased to 15
years and above, and therefore, there was no official data on the incidence of child labour in
1998. Consequently, it was difficult to determine whether the incidence of child labour had
increased during the crisis. Still, it was evident that the crisis had pushed the children of the
poor onto the streets. These children usually worked as beggars or street children, many of
them trying to help their parents make a living. In 1998, according to the Minister of Social
Affairs, the incidence of street children increased by 400 per cent,33 the majority of whom
lived in Jakarta. The number of street children in Jakarta quadrupled following the crisis.

The IFLS data 34 also suggested a significant decline in health services for children after the
crisis. The number of children under five years of age visiting a community health post,
Posyandu (for immunization, growth monitoring, preventive care and other health services)
declined sharply from 47 per cent in 1997 to only 28 per cent in 1998. However, the data also
showed that the immunization programme was not disturbed during the worst time of the
crisis in 1998. While the severity of the crisis was not fully reflected in the 1998 data, the
longer run impact of the crisis on children, in the form of deteriorating quality of health,
education and other forms of human capital, could be serious especially with the slow
recovery process.

The elderly35
There were around 14 million people over 60 years of age in Indonesia in 1998. Most of
them lived in rural areas (66 per cent), and had lower education levels (90 per cent with
primary schooling or less) than the average population, and those who worked were mainly in
the informal sector. Lower education levels combined with informal sector work meant that
their income was generally below the average.

Due to the lack of a national security system, the crisis unfortunately increased the labour
force participation of the elderly from 50.3 per cent in 1997 to 51.8 per cent in 1998. In other
words, the number of the actively working elderly increased by 400,000 during the crisis.

Similar to the general trend in the labour market, the crisis had also driven the elderly out of
the formal sector. In urban areas, the percentage of the elderly who worked in the formal
sector declined from 24.1 per cent in 1997 to 23.5 per cent in 1998. In rural areas, the
proportion declined from 11.3 per cent to 10.3 per cent, which meant that around 90 per cent
of the working elderly in rural areas worked in the informal sector in 1998. The increasing
participation of the elderly in the informal sector was accompanied by an increase in
underemployment among the elderly. From 1997 to 1998, those who worked for less than 35
hours a week increased from 37.1 to 39.4 per cent in urban areas, and from 55.1 to 58.6 per
cent in rural areas.

   Bisnis Indonesia, September 27, 1998
   Kompas, “Anak Jalanan Meningkat 400 Persen”, September 4, 1998.
   As mentioned in Frankenberg, Thomas and Beegle (1999)
   This part is taken from Sigit (1999)


From economic crisis to social problems
The period between December 1997 and January 1998 marked the turning point of the
Indonesian economic crisis, when the crisis moved to a different and worsening path
compared with other Asian countries in crisis. This was marked by a large economic
contraction, significant structural change in the economy, and a drastic decline in real income.

The economic collapse was transmitted to the social sector through several channels - macro-
level adjustments of output structures and labour markets, leading to unemployment,
underemployment and decreases in real wages; and changing patterns of household income
and expenditure and participation in the labour market. These adjustments in household
income and expenditure led to various social problems, such as an increased incidence of
child labour, decreasing nutrition standards and health-related problems, and these
adjustments would eventually in the long run, lead to a lower quality of human capital.

The lack of government spending on social services, along with the failure of social safety net
programmes during the crisis, also worsened the situation. Combined with the political
turmoil, the breakdown in law and order in some areas, and the low credibility of the
government, these social problems have led to various riots and serious ethnic conflicts in
some areas.

The adjustments and emerging trends in the labour market
Both unemployment and underemployment increased during the crisis, but at a smaller rate
compared with previous expectations. The unemployment rate increased from 4.7 per cent in
1997 to 5.5 per cent in 1998, while underemployment increased from 35.8 per cent to 39.1 per

Due to the flexibility of the labour market, the real impact of the crisis was on declining real
wages. Average nominal wages in the formal sector increased by 17 per cent from 1997 to
1998, but as inflation was almost 80 per cent in 1998, real wages declined by 34.1 per cent.
The decrease in real wages was accompanied by a decrease in labour productivity, but at a
slower rate at around 16.6 per cent.

In terms of sectoral allocation, the crisis seriously affected the non-tradable sectors,
particularly the construction, public utilities, financial and domestic-oriented manufacturing
sectors. Employment reduction in the financial sector continued throughout 1999 because of
major restructuring. Agriculture, on the other hand, was responsible for much of the labour
absorption from contracting sectors.

The crisis increased the share of the informal sector, from 63.3 per cent to 65.4 per cent of the
labour market between 1997-98, as people who were laid off in the formal sector turned to
informal activities to seek new jobs. This annual increase during the crisis was much larger
than the annual growth of the informal sector at around 0.5 per cent during the previous

The movement of labour to the informal sector and to the traditional and natural resource
based sectors such as agriculture, forestry and fisheries, was accompanied by the movement

of employment opportunities to the rural areas. Hence, the process of labour market
transformation from the traditional to the modern, informal to formal and rural to urban based
economy was slowed down, and could even be reversed as a result of the crisis.

The crisis also increased labour force participation, especially of women. The greater increase
in the labour force participation of women was both because of greater demand from the
surviving female oriented sectors (in comparison to the contraction of the male dominated
sectors such as construction), and the need for additional sources of income.

The real effect of the crisis was reflected in the decline in purchasing power, the higher labour
turnover rate, and the greater uncertainty in the labour market. Consequently, this led to the
need to work more hours and to find alternative sources of income for the family. The impact
of the crisis was also evident in the higher competition in the informal sector, the increase in
child labour and street children, the increasing number of family workers, especially women,
and the increasing number of workers unprotected by labour laws.

Also, the crisis increased the number of Indonesians working overseas by five times, from
around 250,000 to more than 1,250,000 workers from 1997 to 1998. In addition to providing
alternative employment sources, overseas work was also a valuable source of much-needed
foreign exchange.
The distributional impact and poverty issues
The impact of the crisis on the labour market varied across regions in the country. Although
the data suggested that Java was worst hit by the crisis, the negative effects of the drought
made it difficult to isolate the impact of the crisis in other areas. There was also no correlation
between economic conditions before and after the crisis. Mainly because of the flexibility of
the labour market, the effects on employment did not show significant differences across
regions, even though the decline in real income was relatively larger in Java compared to
other regions.

Reliable data on income are difficult to obtain, but preliminary data on expenditure – with all
the attendant weaknesses in representing income – showed that the crisis reduced the real
expenditure of the wealthy and the highly educated more than the poor and the uneducated.
Therefore, the distribution of expenditure is more even after the crisis.36 The data on wages in
the formal sector consistently indicated that the decrease in real wages for those with higher
education was generally higher than that for those with lower education.

However, the available data would not adequately capture the differing adjustments of rich
and poor households to changes in income and expenditure. While rich households could
dissave to compensate for loss of income, for poor families, the coping mechanisms were
mainly to work longer hours, or to increase the participation of other household members in
the labour market. Therefore, the crisis resulted in an increase in the labour force
participation of women, and an increase in street children working for their parents, although
the data on the impact on vulnerable groups was limited. As a result of the lack of social
security, the elderly were also forced to increase their labour force participation.

The proportion of the population living below the poverty line increased during the crisis, but
less than the initial predictions, which was the subject of controversy. Later estimates
suggested that the number of the population living below the poverty line increased by around
two-thirds in 1998. This being the case, the level of poverty in 1998 was similar to that in the
early 1990s.

  Even though the income distribution of the lower half of the population (below the median income)
could be worse.

Some policy recommendations
The crisis provides pointers for the development of labour policy and ways of mitigating the
impact of a future crisis. One of the most important issues is the need to maintain labour
market flexibility as an essential condition for faster recovery. While there should be no
intervention on wages, and markets should be allowed to run efficiently and make the
necessary adjustments in prices and wages, social safety nets should be provided to protect
workers and other vulnerable groups. The heterogeneous impact of the crisis, suggests that
safety nets should not be targeted just at individuals, but to areas or sectors that are most
severely affected. In other words, geographical targeting is crucial.

Employment creation programmes should facilitate labour mobility from the shrinking sectors
to the booming ones. Instead of preventing lay-offs in contracting sectors, policies should
provide incentives for labour expansion in expanding sectors. Related to this, there is no
crucial need to provide unemployment insurance for displaced workers in the short run, partly
because of the dominance of the informal sector in the labour force, and the lack of
government capacity in handling the insurance programme. The type of jobs offered in the
public works programmes in urban areas should be focused on activities relating to shrinking
sectors such as construction.

In view of the critical role played by the informal sector in labour absorption, the
development of the informal sector and small and medium entrepreneurship should be
encouraged. The establishment of informal markets in certain places in urban areas – such as
“night markets” – would help regulate these informal activities. For small-and-medium
enterprises, an equal access to credit with market-based interest rates would be more efficient
in fostering the development of small businesses than the provision of a heavily subsidized
interest rate for small-scale enterprises.

The flexible nature of the labour market would allow public investment in human capital for
the relatively skilled workers to be reallocated for spending on general education, where the
social returns on investment are higher. Crisis-relief policies should also prevent the
deterioration in the quality of education, through block grants to the poorest schools and an
efficient programme of teacher retraining.

The failure of the social safety net programmes in Indonesia worsened the social impact of the
crisis, and therefore the need to build an efficient and effective safety net is crucial. There are
actually various ways to improve the performance of the safety net programmes. First, social
safety nets should be more decentralized so that local government, local non governmental
organizations (NGOs) and the local community can create their own programmes, distribute
the funds and then monitor the implementation in their localities. In other words, the approach
should be bottom-up rather than top-down.

Second, considering the absence of good governance in terms of capacity, transparency,
control mechanisms and legal framework, it is very important to provide a control mechanism
that minimizes leakage and corruption. In this case, NGOs can play a vital role.

Third, the social safety net programme should not overemphasize the short-run impact of the
crisis such as problems associated with unemployment and income reduction. The flexibility
of the labour market has basically led to the needed short-run adjustment to the crisis. The
data suggested that the longer-term impact of the crisis, such as children’s education and
health related issues, could be more serious. Therefore, the social safety net programmes
should also provide the required support for education and health services in the longer run.

The government should create supporting institutions and policies to facilitate the increasing
number of Indonesians working overseas. It is expected that the number of Indonesians

working in foreign countries will increase after the crisis and beyond, and therefore, support
facilities such as special offices or institutions for training and protecting workers, are needed.
Private sector participation in such measures should be encouraged, both to increase the
efficiency of the programme and to minimize leakages.

Finally, the need for public employment reforms is critical especially when government
revenues are low. One of the major lessons of the crisis was that the lack of capacity of the
government worsened the impact of the crisis. Therefore, reforms such as efforts to downsize
and make the government more efficient are urgently required. Political will is a prerequisite
for the initiation of such measures.

Agrawal, Nisha. "The Benefit of Growth for Indonesian Workers".        World Bank Policy
Research Working Paper, no. 1637. Washington DC. 1996.

Andrew D. Mason and Jacquiline Baptist. “How Important Are Labour Markets to the
Welfare of Indonesia’s Poor?”. World Bank Policy Research Paper, no. 1665. Washington
DC. 1996.

Bisnis Indonesia, Calo Puas Merneras, Pekerja Memelas, November 6, 1998.

Bisnis Indonesia, September 27, 1998

Cameron, Lisa. “Survey of Recent Developments.” Bulletin of the Indonesian Economic
Studies.Vol. 35, No.1. 1999.

Central Board of Statistics. Sakernas (National Labour Force Survey). Various Years.

Feridhanusetyawan, Tubagus. “The Social Impact of the Indonesian Economic Crisis” The
Indonesian Quarterly, Vol 26, no. 4, December 1998.

Frankenberg, Elizabeth, Duncan Thomas, Kathleen Beegle. “The Real Cost of Indonesia’s
Economic Crisis: Preliminary Findings from the Indonesian Family Life Surveys”. Labour
and Population Program Working Paper Series 99-04. Rand Corporation. March 1999.

International Labour Organization (ILO). Employment Challenges of the Indonesian
Economic Crisis. ILO Report. 1998.

Kompas, “Anak Jalanan Meningkat 400 Persen”, September 4, 1998.

Kompas, “1700 per bulan, TKI Masuk Sarawak”, April 23, 1998.

Lee, Eddy. The Asian Financial Crisis: The Challenge of Social Policy. ILO, Geneva. 1998.

Manning, Chris. “Approaching the Turning Point?: Labour Market Change Under Indonesia’s
New Order.” The Developing Economies, Vol. 33. No 1. March 1995.

Manning, Chris. Indonesian Labour in Transition: An East Asian Success Story? Cambridge
University Press. 1998.

Manning, Chris. “Labour Markets in the ASEAN-4 and the NIEs.” Asian Pacific Economic
Literature. Vol 13. No. 1. May 1999.

Manning, Chris and Sisira Jayasuriya. “Survey of Recent Development,” Bulletin of
Indonesian Economic Studies, August 1996.

Poppele, Jessica, Sudarno Sumarto, Lant Pritchett. Social Impact of the Indonesian Economic
Crisis: New Data and Policy Implication. SMERU (Social Monitoring and Early Response
Unit) Report. February 1999.

Republika, “Malaysia Mendeportasi 10.000 TKI Ilegal”, March 21 1998.

Sumarto, Sudarno, Anna Wetterberg, and Lant Pritchett. The Social Impact of the Crisis in
Indonesia: Results from a Nationwide Kecamatan Survey. SMERU (Social Monitoring and
Early Response Unit) Report, 1998.

Suara Pembaruan, Permintaan TKI meningkat, July 11, 1998.

Sigit, Hananto. Social Impact of the Economic Crisis. Unpublished preliminary report for the
Asian Development Bank. May 1999.

The Jakarta Post, 'Steps to protect workers 'slow compared to promotion'", December
18, 1998

UNSFIR. The Social Implications of the Indonesian Economic Crisis: Perceptions and
Policies. Discussion Paper No.1. United Nation Support Facility for Indonesian Recovery
(UNSFIR). April 19, 1999.

World Bank. Indonesia in Crisis: A Macroeconomic Update. World Bank, Washington DC.

World Bank. “Addressing the Social Impact of the Crisis in Indonesia: A Background Note
for the 1998 CGI”. Mimeo. World Bank, Washington D.C. (also available through the
internet). 1998b.


To top