Reading a Falling Unemployment Rate Correctly
A newspaper headline reads: 'Unemployment falls from 6% to 5%, government claims employment miracle.' The figures come from a household survey. Population is 1,400 million; the population aged 15 and above is 1,000 million. In the current survey round, 500 million people are in the labour force and 475 million are employed. The previous year's revised estimates show a labour force participation rate of 52% and an unemployment rate of 6%. Separately, 45% of those employed work in agriculture, a sector that produces about 15% of gross value added, and about 90% of workers are in informal employment. An opposition member asks how employment can have 'fallen' when unemployment 'fell'. The statistical office notes that the survey also offers a current weekly status measure, which gives a higher unemployment rate than the usual status measure used in the headline.
1. Compute the labour force participation rate, worker population ratio and unemployment rate, and reconstruct last year's figures. Does the data support the headline?
Read the worked answer
Current round:
- LFPR = labour force / population aged 15+ = 500/1,000 = 50%
- WPR = employed / population aged 15+ = 475/1,000 = 47.5%
- Unemployment rate = unemployed / labour force = 25/500 = 5%
Last year:
- Labour force = 0.52 × 1,000 = 520 million
- Unemployed = 0.06 × 520 = 31.2 million
- Employed = 520 - 31.2 = 488.8 million; WPR = 48.88%
Comparison: the unemployment rate fell from 6% to 5%, but employment fell from 488.8 to 475 million - a loss of 13.8 million jobs - and the labour force shrank by 20 million.
Judgement: the headline is misleading. The unemployment rate is a ratio whose denominator shrank: 20 million people left the labour force (discouraged workers, students staying in education, women withdrawing), and that is what pulled the rate down. The accurate statement is that employment fell and participation fell; the 'miracle' is a denominator effect, not job creation.
2. Which indicator should be used to judge 'improvement', and how does the agriculture data complicate the picture?
Read the worked answer
No single ratio is sufficient. Use a small dashboard:
- Employed persons (level): did the number of people with work rise?
- WPR: what share of the working-age population has work?
- LFPR: are people participating or withdrawing?
- UR: how much slack exists among those who want work?
For welfare, add real wages and the share of workers in low-productivity work. The UR alone can improve simply because people stop looking, as this case shows.
Agriculture: 45% of the employed (about 214 million) produce 15% of GVA, so value added per agricultural worker is roughly one-third of the national average. This pattern is consistent with disguised unemployment: output would not fall much if some workers left the land. A rise in 'employment' that consists of more people sharing the same farm work is therefore not an improvement in welfare.
Caveats: usual status (reference period of a year) and current weekly status (one week) give different rates, the CWS rate being higher because it captures short spells; 90% informality means earnings and underemployment are poorly captured; and a 1 percentage-point change may be within sampling error.
3. Draft two sentences of guidance to the newspaper on how to report the figure and name one additional statistic that would settle the question.
Read the worked answer
Suggested wording: 'The unemployment rate fell from 6% to 5%, but the number of employed persons fell by 13.8 million and labour force participation fell from 52% to 50%, so the fall in unemployment reflects withdrawal from the labour force rather than job creation.' 'The survey also publishes a current weekly status measure, which gives a higher unemployment rate than the usual status measure quoted here, and readers should be told which is being used.'
Why this is the right framing: the unemployment rate is unemployed divided by the labour force. A shrinking numerator and a shrinking denominator can produce a fall in the ratio even as employment declines. Reporting all three numbers lets the reader see which effect is driving the change.
Additional statistic: growth in real wages of regular workers (alternatively, the sectoral split of the change in employment, or the CWS unemployment rate). If real wages rose while employment fell, the story would be one of rising productivity rather than distress; without wage data, the withdrawal could be either.