§ 1998. Mr. Gibb
To ask the Secretary of State for Work and Pensions (1) how many prosecutions for social security fraud there were in(a) 1997, (b) 1998,(c) 1999 and (d) 2000; 
(2) how many claimants were found to be illegally claiming social security benefit for a year or more in (a) 1997, (b)(c) 1999 and (d) 2000. 
§ Malcolm Wicks
Information on the duration of fraudulent claims is not collected centrally and could be obtained only at disproportionate cost. In order to prepare for the implementation of the "two strike" provisions in the Social Security Fraud Act 2001 we are currently improving the data sources used to compile information on prosecutions, cautions and penalties. The latest provisional information is in the table.
Investigations Prosecutions Cautions and penalties as an alternative to prosecution Total Benefits Agency 1997–98 11,700 — 11,700 1998–99 10,238 2,835 13,073 1999–2000 9,124 11,029 20,153 2000–01 11,403 15,555 26,958 Local authority 1997–98 700 — 700 1998–99 800 — 800 1999–2000 860 390 1,250 2000–01 1,100 550 1,650
1. Benefits Agency investigations data are now compiled from a new data source. The figures above are provisional and may differ from those previously published.927W
2. The data on cautions and penalties for 1998–99 is derived from the database proceeding the current database and there maybe some overlap with 1999–2000.
3. Administrative Penalties, as an alternative to prosecution, were introduced by the Social Security Administration (Fraud) Act (1997) with effect from December 1998.
4. Figures are taken from local authority management information returns. It is possible that there could be some double counting with Benefits Agency data if there were cases which involved a joint prosecution.
5. Data are not available for all 409 local authorities. The total for Great Britain includes estimates for local authorities that have not responded. These estimates are based on historical and regional data. This type of estimate is standard practice in reporting totals where there have been non-respondents. The figures have been rounded to two significant figures.