Hostname: page-component-78c5997874-v9fdk Total loading time: 0 Render date: 2024-11-15T08:57:25.949Z Has data issue: false hasContentIssue false
  • English
  • Français

The Measurement and Behaviour of the UK Saving Ratio in the 1970s

Published online by Cambridge University Press:  26 March 2020

K. Cuthbertson
Get access Rights & Permissions [Opens in a new window]

Extract

The consumption function is probably the most widely researched area in macro-economics and until the early seventies it seemed to be one of the most stable relationships. However, in the 1970s there was a sharp and rapid rise in the saving ratio of the personal sector that was not predicted by existing equations. Moreover the rise in the saving ratio coincided with an increase in the rate of inflation and the conventional view had been that inflation would discourage saving. This prompted a search for theories that would explain the ‘paradox’ of saving behaviour.

Type
Articles
Information
National Institute Economic Review , Volume 99 , February 1982 , pp. 75 - 84
Copyright
Copyright © 1982 National Institute of Economic and Social Research

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)

References

(1) See, for example, Financial Statement & Budget Report 1980/81, p. 27, and ‘The 1981 Budget and the Government's Expenditure Plans 1981-82 to 1983-4, Vol. II, Session 1980-81', HMSO, April 1981, p. 10.

(2) See, for example, C. Bean, ‘The determination of con sumers’ expenditure in the UK', Treasury Working Paper, no. 4, 1978; Davidson et. al., ‘Econometric modelling of the aggregate time-series relationship between consumers’ expenditure and income in the UK’, Economic Journal, vol. 88, no. 4, 1978; D. Hendry and T. von Ungern Sternberg, ‘Liquidity and inflation effects on consumers’ expenditure’, forthcoming in Essays in the theory and measurement of consumers' behaviour, edited by A. S. Deaton.

(3) We therefore build on the work of A. G. Armstrong, ‘An analysis of personal savings in the UK’, The Manchester School, December 1979 and C. T. Taylor and A. R. Threadgold, ‘Real’ national savings and its sectoral composition', Bank of England Discussion Paper, no. 6, October 1979.

(1) These figures are taken from the income of the ‘household sector’ in the NIBB 1981 T4.4. A detailed account of the method of obtaining these figures is to be found in ‘The household sector’ by T. Jones, Economic Trends, September 1981.

(2) We do not deduct a ‘sinking fund’ for the depreciation of dwellings of the personal sector since we think it unlikely that persons undertake such a calculation. Note, however, that expenditure on housing maintenance is included in consumers' expenditure and therefore appears as negative saving of households.

(3) We have retained ‘income in kind’ in our definition of income in order to remain as close as possible to the NIBB 1981 definition of income. Income in kind is part of con sumption and therefore does not affect the level of saving. Changes in the proportion of income in kind in total income do influence the (denominator of the) saving ratio; however such changes are small over the period 1970-80 and do not inhuence our results.

(4) Our measure of ‘net income’ differs from ‘household dis posable income’ in NIBB 1981, T4.4 because the latter treats interest payments as expenditure and also deducts employees' contributions to LAF.

(5) In earlier editions of the NIBB consumers' expenditure by ‘households’ is not published, although it can be calculated using figures from a number of different tables in the NIBB.

(6) ‘Committed and discretionary saving’ by S. Toland, Economic Trends, November 1981.

(7) The figures in S. Toland op. cit. include employers' contributions and the rent interest and dividends earned by LAF net of administrative costs in ‘committed saving’, as well as regular repayments of hire purchase.

(8) If (S/Y)' is the saving ratio excluding stock appreciation, Z, (or the addition to tax reserves or depreciation) and SlY is (the CSO measure of) the savings ratio including such items. then (S/Y)'=(S/Y)(1-Z/S)(1-Z/Y)-1 and Δln(S/Y)≃ Δln(S/Y)‘—Δln(1-Z/S) since Δln(1-Z/Y) << Δln(1-Z/S). Thus (S/Y) rises faster than (S/Y)’ whenever Z increases as a proportion of S. Stock appreciation as a percentage of personal sector saving increased from 3 per cent in 1970 to 6.4 per cent in 1973. The comparable figures for additions to tax reserves are 2.3 and 8.6.

(1) Here we are making the not unreasonable assumption that most of the ‘regular’ savings in LAF, SAYE and mortgage repayments are contractual and can only be reduced (if at all) at a relatively high cost.

(2) This figure overestimates the ‘true’ level of discretionary saving because our definition includes regular repayments of principal on bank loans and hire purchase as noted earlier.

(1) The impact of LAF on households' committed saving is about half of that found by Armstrong op. cit., table 4, because his LAF figures also included single premium payments to LAP and these increased sharply between 1970 and 1972.

(2) ‘Saving in LAF’ by the personal sector as measured by the CSO includes employers' as well as employees' contributions plus rent interest and dividends earned by LAF net of adminis trative costs and payment of benefits (see S. Toland, op. cit. for further details) and is therefore very different from ‘com mitted saving in LAF’ as used in this note.

(3) See table 2.

(4) Studies that use consumption as the dependent variable and the CSOs definition of real personal disposable income (which includes employees' and employers' contributions to LAF and net property income of LAF) and inflation as independent variables may also give biased estimates of the ‘inflation effect’ on saving over the 1970-72 period. This source of misspecification in the consumption (or savings) function may also go some way towards explaining the rather wide range of estimates of the ‘inflation effect’ both in different studies and over different estimation periods.

(5) The exception being a study using time-series data by A. R. Threadgold, ‘Personal savings and the impact of life assurance and pension funds’, Bank of England Discussion Paper, no. 1, 1978. For a recent study using cross-section data see F. Green, ‘The effect of occupational pension schemes on saving in the United Kingdom: A test of the life cycle hypothesis’, Economic Journal, vol. 91, March 1981.

(1) See HM Treasury, Macroeconomic Model: Listing of equations and variable definitions, September, 1981, C. Bean, Davidson et al. and D. Hendry and T. von Ungern Sternberg, op. cit.

(1) This measure of the real return to home ownership is a simplification but probably reflects how the average home owner calculates the return. The real return may be calculated as follows. If PH=average price of housing, K=real stock of housing, and P=consumer price index (the ‘numeraire’) then the real value of the housing stock at time t-1 is R Vt -1= (PH.K/P)t- 1. The change in the real value of a fixed stock of housing (i.e. Kt=Kt- 1) is ΔRVt=RVt-RVt -1=(PH.K)t -1 (PH—P)/Pt where a dot over a variable indicates a percentage change in the variable. The percentage change in the real value is Pt-1(PH-P)/Pt and our measure therefore assumes Pt -1/Pt≃1. We include a measure of the inflation gain/loss on the outstanding stock of mortgage debt later in the article.

(2) Notes and coin, bank and building society deposits, national savings and ‘other short term assets’ as defined in Financial Statistics.

(1) This follows the work of C. Taylor and A. R. Threadgold op. cit., who derive the inflation adjusted savings of the personal sector. We are interested in the saving of house holds and we extend their analysis by looking at the inflation losses on real as well as financial assets.

(2) If the stock of net liquid assets (i.e. liquid assets minus bank advances and hire purchase debt and the outstanding stock of mortgage debt) at the period t—1 is Lt- 1, R =average rate of interest on liquid assets, and P=consumer price index, then the change in the real value of a given stock of liquid assets is ΔRV=Δ(L/P)t=Lt- 1 (1+R)/Pt-(L/P)t -1=(R-P) Lt-1/Pt. Hence the ‘inflation loss’ in current period prices is (PLt -1) where P=rate of inflation in consumer prices. Note that the inflation loss/gain on the outstanding stock of mort gage debt is included in these figures, and we are treating the gains and losses on assets and liabilities equivalently: indivi duals may of course not do so. Finally, note that the inflation loss on assets is large, and greater than the inflation gain on liabilities thus giving rise to the relatively small loss on net liquid assets shown in table 3.

(3) It could be argued that a substantial proportion of the saving of persons is for the purpose of house purchase. In this case the inflation loss would be greater than the figures in table 4 when house prices rose faster than the CPI (e.g. 1970-73) and vice versa (as for example between 1974 and 1977).

(1) Defined as the net discretionary saving of households (column 3, table 3) as a percentage of the net income of house holds (appendix table 1).

(2) Reducing balance method of depreciation. medium ‘length of life’. See Economic Trends, March 1978 for further details.

(3) This we have again measured as the net stock of dwellings of persons multiplied by the difference between the rate of inflation in house prices and the consumer price index. The inflation gain on the outstanding stock of mortgage debt is included in the inflation loss on liquid assets (see above foot notes on page 81). Even if only a fraction, say 1 /10th, of home owners are planning to realise any inflation gain on housing or take cognizance of the gain, the figures are still substantial in relation to, say, the inflation loss on net liquid assets.

(4) See K. Cuthbertson, ‘The determination of expenditure on consumer durables’, National Institute Economic Review, no. 94, November 1980 for evidence that durables expenditure is positively related to the rate of inflation.

(5) See, for example, A. S. Deaton, ‘Wealth effects on con sumption in a modified life cycle mode’, Review of Economic Studies, vol. 32, no. 120, 1972 and the pioneering work of A. Ando and F. Modigliani, ‘The Life Cycle’ hypothesis of saving: aggregate implications and tests', American Economic Review, vol. 53, no. 1, part 1, 1963.

(6) In the regression results cited above ‘errors’ in 1978 and 1979 are small relative to those for 1980.

(7) Rising interest rates and capital losses in 1979 on market able assets might also influence saving, although such losses might have been expected to be reversed in 1980 and indeed interest rates did fall through 1980.

(1) Of course inclusion of the rate of inflation as an additional variable might be indicated by alternative theories to the one advanced here. See, for example, the ‘rational money illusion hypothesis’ of Deaton, 1977, op. cit. and G. Bulkley, ‘Personal Savings and Anticipated Inflation’, Economic Journal, vol. 91, no. 136, March 1981. However, D. Hendry and T. von Ungern Sternberg and the National Institute of Economic and Social Research, op. cit. find that the rate of indation is statistically insignificant when included with terms in the inflation loss in liquid assets.

(2) Since portfolio theory leads one to expect a fall in the liquid assets to income ratio (L/Y) at higher rates of inflation, the inflation effect (β.L/Y)P would be smaller than the figures given here. We require a separate model of the determinants of (L/Y) to estimate this additional influence.

(3) Although it is argued in the Financial Statement and Budget Report 1980/81 (p. 27) that the inflation effect works by ‘influencing’ the amount of saving necessary to maintain intact the real value of financial assets, the Treasury's con sumption function does not embody this mechanism explicitly (see table 3, C. Bean, and HM Treasury, op. cit.). Our calculations of the inflation elasticities assume a zero rate of growth of real income: at higher rates of growth of real income the inflation elasticities would be smaller.

(4) First, it is also worth noting the variability in the estimates of the response of saving to a change in the rate of growth of income in the econometric studies of table 2: D. Currie, ‘Some long run features of dynamic time series models’. Economic Journal, vol. 91, September 1981 suggests that these effects should be tested more rigorously than hitherto. Second, HM Treasury's (non-durable) consumption function contains a term in unemployment which reduces consumption (in creases ‘precautionary saving’) as unemployment accelerates. In contrast, T. von Ungern Sternberg, ‘Inflation and savings: international evidence on inflation-induced income losses’, Economic Journal, vol. 91, December 1981 reports that he failed to find any unemployment effects in a consumption function that also included the inflation loss in liquid assets.

(5) See S. Toland and T. Jones op. cit.