In the Labour Force Survey (LFS) respondents are interviewed for 5 consecutive quarters over a 12 month period, with 20% of the sample being replaced at each quarter. This allows for a longitudinal dataset to be created over a limited time interval, where respondents’ characteristics can be tracked over their time in the survey.
We publish population-weighted longitudinal datasets for each calendar quarter. These are available for each quarter since 1997 and can be used to analyse changes in labour market characteristics over 2 or 5 quarters. The datasets include "flow" variables, which estimate the size of the movements between the 3 main labour market statuses of employment, unemployment and economic inactivity.
Monitoring changes in the labour market status of respondents to the LFS aids the understanding of the quarterly changes in the levels of employment, unemployment and economic inactivity. These indicators are published as stocks for a given period, with changes expressed as the difference between successive quarters. These quarterly comparisons represent the net changes between the 3 labour market statuses. The underlying gross flows are usually considerably larger and may not correspond with those implied by the net changes. Estimates of the gross flows between the statuses can be derived from the LFS Longitudinal Datasets and are summarised in this note.Nôl i'r tabl cynnwys
There are 2 types of LFS longitudinal datasets: 2-quarter and 5-quarter. These are weighted using the same population estimates as those used in the main quarterly LFS datasets, although the weighting methodology differs (see technical note). Consequently the estimates are broadly consistent with the published aggregates, but not entirely. Also, the datasets are limited to people aged 16 to 64.
Both types of dataset contain a flow variable with 11 categories, with all combinations of employment, unemployment and economic inactivity accounted for, plus 2 categories for those entering and leaving the 16 to 64 population over the quarter. For the purpose of this analysis, those entering or leaving this population are excluded from the measured sample. The stock of the employed, unemployed and inactive at each quarter can therefore be estimated by summing the corresponding flow categories.
For this analysis, the 2-quarter datasets have been used in order to gain some insight into the quarterly changes in the headline published aggregates.Nôl i'r tabl cynnwys
The charts in this article show the estimated gross flows, that is the total inflow or outflow for 16 to 64 employment, unemployment and inactivity from 1 calendar quarter to the next. They are seasonally adjusted. Analysis of the net flows, that is, the difference between the total inflow and outflow, are also included and these are compared with the quarterly changes in the published aggregates, partly to give an indication of the robustness of the flows analysis.Nôl i'r tabl cynnwys
The flow from employment to unemployment has increased but remains under 800,000.
The gross flow into employment has increased to its highest level since comparable records began in 2001.
The flow into employment from inactivity is at its highest since October to December 2004.Nôl i'r tabl cynnwys
The diagram shows the gross flow between each economic status between July to September 2015 and October to December 2015. The stocks for each status represent the latter period and are the seasonally adjusted aggregates for people aged 16 to 64.
Nôl i'r tabl cynnwys
An increase in the flow from employment to unemployment has caused the gross flow to unemployment to increase following a decrease last quarter (Figure 1).
The flows to employment and inactivity (Figure 2) have shown little change over the year.
Figure 3 shows that the quarterly change in stock and net flow have both increased but remain negative.
Nôl i'r tabl cynnwys
The flow to employment from inactivity (Figure 4) has increased. The gross flow into employment has increased to its highest level since comparable records began in 2001.
The flow from employment to inactivity (Figure 5) is flat whereas the flow to unemployment increased after a decrease last quarter.
The job to job flow (Figure 6), that is, the number of people who remained in employment over the quarter but are in a different job, has decreased over the quarter.
Figure 7 shows that the net flow has decreased whereas the quarterly change in stock has continued to increase.
Nôl i'r tabl cynnwys
The flows from both employment and unemployment into inactivity have shown little change over the last 4 quarters and the gross flow remains flat (Figure 8).
The gross flow from inactivity (Figure 9) is unchanged on the quarter with the flows to employment increasing and unemployment decreasing. The flow to employment is at its highest since October to December 2004.
Figure 10 indicates that the quarterly change in stock and the net flow have decreased and are both negative.
Nôl i'r tabl cynnwys
There are differences between the data used for the published LFS aggregate estimates and the longitudinal data used to estimate the gross flows:
Flows are currently adjusted for non-response bias through special calibration weights in the longitudinal datasets. These aim to account for the propensity of certain types of people to drop out of the LFS between one quarter and the next. For example, housing tenure features in the weighting of the longitudinal data because, historically, households in rented accommodation have been more likely to drop out of the survey than owner-occupiers.
There is some evidence that the longitudinal datasets are affected slightly by response error which causes a slight upward bias in the estimates of the gross flows. For example, if it was erroneously reported that someone had moved from unemployment to employment then, in addition to the outflow from unemployment being overestimated, so would the inflow to employment. In the main quarterly LFS dataset, any such misreporting errors tend to cancel each other out.
The differences in the net flows for inactivity shown in Figure 10 are mainly the result of excluding the entrants to, and leavers from, the population in the flows estimates contained in this piece of analysis. This effect is normally one that increases the number of people who enter inactivity. This is because the increase in inactivity from those people turning 16 is greater than those leaving inactivity due to becoming 65.
The stocks derived from the longitudinal datasets differ from those obtained from the quarterly LFS datasets due to their being based on a subset of the main LFS sample. The restriction to measuring only those who are commonly aged 16 to 64 across successive quarters discounts those entering or leaving the population and also those over 64. All such people are accounted for in the headline LFS aggregates.
Manylion cyswllt ar gyfer y Erthygl
Ffôn: +44 (0)1633 455400