One of the most striking things one first notices when examining the survey data is that cultivation costs rose by 14.6% over the period 2002 to 2004.

Labour costs
The second more important reason that cultivation costs rose was the major rise in labour costs (+ Rs. 5959, 22.8%). This rise is due almost entirely to the rise in cross-pollination costs – where most child labour is used. Cross-pollination (per acre per season) costs rose from Rs. 23,306 to Rs. 29,239 (+ Rs. 5934 or a 25.5% rise over the 3 years). Thus cross-pollination accounts for 99.6% of this rise in labour costs and it is also the activity where the most child labour is used.
What distinguishes cottonseed production from other local crops - such as paddy, jowar or cotton - is that the production of hybrid cotton seed is very labour intensive. A chief part of this production is cross-pollination which is done manually and this activity alone requires about 90.3% of the total labour expended and 45% of the total capital used and is done mostly by female children. Cross-pollination involves two separate activities: emasculation and pollination. The buds, which are expected to open up the next morning, are selected for 'emasculation'. Emasculation should then be done in the afternoon. These buds should be gently removed without injuring the gynaecium. Doak's method of emasculation of the flower bud is used, which consists in the removal of bracts first by hand, and then the petals, along with the entire anther-sac whorl, with the nail of the thumb, without damaging the stigma, style or ovary. The emasculated flower should be carefully covered with red plastic paper bags. Pollination should be done in the morning hours with the pollen of male parents. Crossing needs to be done as soon as the flowers blossom before the female flowers bear fruit (and consequently produce non-hybridised or 'fake' seeds). It is clear from the above discussion that cross-pollination work needs to be done very carefully, with much patience, by a disciplined work force. About 2 months after sowing, the plant starts blossoming and continues growing for five months. During this time, cross-pollination (both emasculation/pollination) need to be done everyday, without fail. The duration of cross-pollination is three to four months.

Wage inflation

Why have cross-pollination costs risen so much – 25.5% - over the past three years? One reason is that average wages in cottonseed production have risen by 16%.

However, an even more direct influence on rising wages and on the rise cross-pollination costs in cottonseed is the result of some replacement of child labour with more expensive adult labour. The estimates presented in Table 4 below indicate the daily use of labour for cross-pollination activity for one acre. We find that the use of adult hired labour has nearly doubled since 2002, rising from 1.74 to 3.66 (+1.92; a 110% rise) persons per acre since 2002. Adult hired labour has increased mostly at the expense of child hired labour (-1.33; or 24% fall) and partly at the expense of a fall in family labour (both child and adult, the combined fall is -0.13; or 13.1% fall). Together these three categories (child hired labour, adult family labour and child family labour) account for a fall of 1.46. So how is it possible that adult labour has risen 1.92 person per acre? This rise is the result of a rise in total labour needed overall (+.46 per acre).
Thus in sum, the substitution of child and family labour for more adult paid labour has:

1. pushed up average wages because adults are paid more than children;
2. increased total labour required because children and family labourers work longer hours per day than adult labourers do.

1. Wage inflation: wages in cottonseed production have increased by 16% in line with inflationary wage increases in agricultural field work outside of cottonseed production, and
2. Workforce restructuring: there has been a substitution of child workers and family workers for paid adult labour. This substitution has increased wages further as well as increased total labour required.

If one assumes that 16% of the rise in wages would have occurred regardless of the substitution, then one could estimate that the remaining 9.5% (25.5%-16%) was the direct result of the substitution of adult for child labour. Thus farmers spent an average of Rs. 5959 per acre more on cross-pollination costs in 2004 than they did in 2002, of which Rs. 3746 is due to 16% increase wages and remaining 9.5% or Rs. 2213 is the direct result of substitution of adult for child labour.
Whilst this is a useful estimate, we think this underestimates the true cost of the substitution, because wage rises outside of agriculture are likely to be much more responsive to supply and demand pressures than within cottonseed production, where the wage rises are marked by more stickiness upwards⁷. Hence some of what we attributed to wage inflation (16% rise) may instead be a product of this substitution.

Teenagers and Low Hourly Wages

This retrospective estimation does not give a good indication of the comprehensive costs of replacing child with adult labour because in 2004 approximately 70% of the 'adults' are in fact teenagers between 15 and 18 years. The remaining 30% are adults over 18.
Teenagers in the age group of 15 to 18 years were categorized as ‘adults’ by our farmer employers respondents who wished to subvert child labour laws by claiming to employ adults at what appears to be adult daily/monthly wages. They were so convincing that even our field investigators categorized these teenagers as adults. However, hourly wages of teenagers were indeed lower than adults over 18 because the former regularly worked longer hours and were found to be subjected to the same type of control that children worked under, including increased work intensity, verbal and physical abuse and exposure to pesticides. These oppressive conditions are designed to get a higher work output for less payment. Adults over 18 - with access to outside higher wage opportunities - would not accept and indeed do not put up with such conditions (see D. Venkateswarlu and L. da Corta, 2001)⁸. Most of these teenagers worked for these employers when they were younger – the older child group fell by 7.9% in line with the rise in teenagers by 7.9%, see Table 5 below.

Unfortunately, the upward pressure on cottonseed wages through the inclusion of adults did little to equalise average wages in cottonseed (an average of the child, teenage and adult wages) with outside wages for adult men and women in other agricultural work: even in 2004 average cottonseed wages were below wages in other agricultural activities.

To sum up thus far, cultivation costs rose 14.6% over the period (2002-4) attributable partly to the rise in input costs (rent, pesticides) but also largely due to the rise in labour costs, chiefly cross-pollination costs, which have risen 25.5% over the period. Cross-pollination costs rose as a result of the substitution of paid child labour and family labour for paid adult labour. Using hourly wages, we showed the impact of this substitution was not merely in increasing wages, but because paid adults work much shorter days than children (and family workers), it has increased labour required which further increased labour costs. We estimated that out of the Rs. 5959 average cost rise in cross-pollination costs per acre, Rs. 3746 is due to the 16% increase in wages and remaining 9.5% or Rs. 2213 is the direct result of substitution of adult for child labour.
We also argued that before getting too optimistic about these early stages of the switch to adult labour, one must bear in mind several issues:

1. 70% of ‘adult’ labour in cottonseed production are teenage (15-18 year old) workers who were probably kept on from when they were younger and who continue to work longer hours for a lower hourly wage than those adults over 18.
2. The upward pressure on cottonseed wages through the substitution of child for adult labour did little to equalize average wages in cottonseed with outside wages: even in 2004 cottonseed wages were below wages in other agricultural activities¹⁰.

Table 10 shows that per acre income rose slightly from Rs. 78,774 in 2002 to 80,495 in 2003 (Rs. 1721 or +2.2%) but then fell dramatically in 2004 (from Rs. 80,495 in 2003 to Rs. 64,366 in 2004 or -20.0%) as a result of a pest attack which reduced output.

Though farmers have what appear on the surface to be a reasonable profit in 2002, over the three year period profits have declined attributable to a threefold assault on profits: rising costs, falling or variable output, and low or falling buying prices. Firstly, costs have risen, both non-labour costs (like rent and inputs such as pesticides, fertilizers and interest) and labour costs. Labour costs rose way above inflationary wage rises as a consequence of the pressure to substitute child with adult labour. Secondly, this rise in costs took place within the context of variable and falling output due to periodic pest attacks. Thirdly, buying prices have fallen or not kept up with inflation: the buying price of lint fell 30% and the buying price of commercial cotton fell 6.5% since 2002. And whilst the seed price rose 8.5%, it was only ½ of wage inflation (16.4%) and far below the 25.5% rise in cross-pollination costs. All these factors, coupled with debt racked up during bad years, put pressure on farmers to take a very conservative view on substituting cheap child for more expensive adult labour. It also encourages them to flout labour laws by hiring children and teenagers who work longer hours, work more intensively and who are easier to discipline and to control than adults over 18. Smaller farmers, despite their higher output in each year due to greater intensity, were crushed by higher costs, especially interest payments, and this showed up in their profits which were lower than big farmers.

In sum, over the period 2002-04, we have identified:

1. Rising labour and non-labour input costs,
2. Fluctuating output,
3. Falling or slow rising prices for lint, cotton and cottonseed.

SECTION-I: Introduction

SECTION-III: Cost implications of replacing child labour