Question 1.PR.13: The following data show car registrations in the UK during 1...
The following data show car registrations in the UK during 1970–91 (source: ETAS, 1993, p. 57):
| Year | Registrations | Year | Registrations | Year | Registrations |
| 1970 | 91.4 | 1978 | 131.6 | 1986 | 156.9 |
| 1971 | 108.5 | 1979 | 142.1 | 1987 | 168.0 |
| 1972 | 177.6 | 1980 | 126.6 | 1988 | 184.2 |
| 1973 | 137.3 | 1981 | 124.5 | 1989 | 192.1 |
| 1974 | 102.8 | 1982 | 132.1 | 1990 | 167.1 |
| 1975 | 98.6 | 1983 | 150.5 | 1991 | 133.3 |
| 1976 | 106.5 | 1984 | 146.6 | – | – |
| 1977 | 109.4 | 1985 | 153.5 | – | – |
(a) Draw a time-series graph of car registrations. Comment upon the main features of the series.
(b) Draw time-series graphs of the change in registrations, the (natural) log of registrations, and the change in the ln. Comment upon the results.
(a) Strong upwards trend for 1974–1989, but substantial falls in 1972–1974 and 1989–1991. The market appears quite volatile, therefore. Note that this shows the volume of car registrations. If the 1972–1976 pattern is repeated, one might expect car registrations to turn up again in 1992–1993.
(b) This figure shows the substantial volatility of the series, around an average not much above zero. 1989–1991 looks less like 1971–1973 on this graph. The log graphs are very similar to the levels graphs. There is not always an advantage to drawing these. Here, the time period is relatively short and the growth rate rather small.