the increase in the working-class cost-of-living since before the war
TRANSCRIPT
THE INCREASE IN THE WORKING-CLASS COST-OF-LIVINGSINCE BEFORE THE WAR
The recently-issued Supplement No. z to the Industrial Relations Hand-book describes in detail the structure of the new Index of Retail Pricesincluding particulars of the weighting system, which is based on the repre-sentative working-class household budgets collected by the Ministry ofLabour in i7 and 1938.' The weights' represent the proportions shown bythe 1937-8 budgets, adjusted to take account of the broad changes in relativeprices between that date and mid-June, 1947.' The Supplement adds: 'Theinformation available on relative price changes since 1937-8 was sufficientto provide, for weighting purposes, broad estimates of the proportions inwhich the various sections and groups entered into the total cost at the basedate. It was not, however, sufficiently accurate to compare the actual cost ofthe items in 1937-8 and at the base date, and hence, to give an index of pricechanges since pre-war.' Now this question of the change since before thewar in the working-class cost-of-living, in the sense of the cost of buying whatwas bought by representative working-class families in 1937-8, is central foreconomic discussion. While such a calculation would only be an approxima-tion to the increase in the cost of buying what the working-class purchased in1937-8, and while this increase is not entirely relevant to the more funda-mental question of the increase in the cost of maintaining a certain livingstandard, an estimate of the order of magnitude of this change is essentialfor such questions as the redistribution of real income since pre-war, andthe change in the purchasing power of wages and social security benefits.The major changes caused by the war have probably by now already takenplace, and if the experience after the 1914-18 war is a guide,' there should bea new plateau in the cost of living established, variations over which will besmall compared with the change since before the war.3 The rough differencebetween this plateau and the level before 1939 is a fact of major importancein discussions of the larger economic and political problems. Technicalmodesty would not seem by itself an adequate reason for suppressing theCommittee's estimate, even if ' broad', and perhaps a more weighty reasonwas an unwillingness for such an estimate to be compared with the rise in thecost-of-living shown by the old cost-of-living index. Perusal of the Supple-ment indicates that the Ministry has in fact been at some pains to calculatethe estimate with care.
It is possible, however, to reconstruct. by inference the TechnicalCommittee's probable procedure and to estimate their estimate. The prin-ciple on which the Committee worked must have been approximately as
'These budgets were collected for four separate weeks, in October 1937, January 1938,April 1938, and July 1938. When' 1937-8 ' is used in this paper it refers to these dates.
Jgnoring the immediate post-war inflation which was responsible for increases of 20and 30 points per annum, and then decreases of the same order, since repetition of thisinflation has been partly avoided.
'Though this depends on subsidy policy and on the trend of world prices for primaryproducts, which in turn depends largely on the trend of effective demand in the UnitedStates.
THE INCREASE IN THE WORKING-CLASS COST-OF-LIVING 141
follows: The weight Wr for any particular item, taken from the 1937-8budgets (and scaled up so that E(w) = 1,000) must have been multiplied by theprice change (Pr) between 1937-8 and mid-June ¡947 to give a total weight(WrPr) for the base date of the index. This weight had then to be reduced toW, so that (w) = i ,000. The total weights at the base date before reductionmust have been E(wp), and each weight must have been multiplied by i000
for (w) to be i ,000. For each section of the index therefore
WrWrPrZ(wp)¡000 (I)
But since E(w)= i000, (wp) is in fact the CoThmittee's estimate of the¡000
weighted average overall increase between 1937-8 and June ¡947 in the costof purchasing the goods in the representative budget, an estimate we shallcall 'I'. For each section of the index, therefore:
Wr Pwr - I
Verbally, the weight given to each section at the base date (June ¡947) standsin the same proportion to the ¡937-8 weight as the relative price increase forthat section does to the overall weighted average increase in the cost-of-living.'Re-arranging, we have
IP,.- (2)W,
We know the June 1947 weights for each section of the index (i.e. Wr),because these are printed in Appendix C to the Supplement. We can estimatethe 1937-8 weight for each section of the index (i.e. Wr) with considerableaccuracy, because the method used by the Technical Committee is describedin detail in the Supplement. The full reconstruction of the ¡937-8 weights isdescribed in Appendix A, the first of the technical appendices of this paper.If, therefore, we can estimate the price-relative used by the Technical Com-mittee for any section of the index (pr), we can find out from formula (2) theTechnical Committee's estimate of I, the overall weighted average increase inthe working-class cost-of-living (defined as the cost of buying what thebudgets showed the working-class bought in ¡937-8).
There are two sets of price relatives for consumer goods available, therelatives given in the National Income White Paper,2 and those given in theMinistry of Labour Gazette for sections of the old cost-of-living index.Each price relative can be used in formula (2) to give us a value of lit implies.
'The argument does not, of course, depend on the assumption that the Committeechose weights so that I (w,) = 1,000. 1f the Committee did not, then
wrL E(w)The expression inside the bracket is in fact w,, scaled up so that Z (w)=- 1,000. Whateverthe Committee's I (w) was, therefore, we can treat w, in equation (2) as if I (w) 1,000.
Cmd.7371
142 THE BULLETIN
If these relatives were the ones used by the Technical Committee, everyimplied estimate of I would be the same, except for small differences of apoint or two due to rounding. In fact, however, some of these relatives wereinappropriate for estimating the increased cost of working-class purchases ofthe commodity groups concerned, as revealed by the 1937-8 budgets. The pricerelatives in the National Income White Paper refer to total consumer purchasesof each group of commodities. They are likely to be particularly inappro-priate where the prices of goods bought by the working class have increasedless than a group's average price, because of Governmental policies of sub-sidising necessities and subjecting them to more rigorous price controls.In such cases p will be too high, if we assume, as we must, that the TechnicalCommittee would avoid such biassed relatives. The implied value of I willtherefore also be too high. The price relatives for items in the cost-of-livingindex are in many cases more likely to have been used by the Committee.They were readily available, and most of them referred to homogeneouscommodities which were not subject to a discriminating price policy. Thecost-of-living index price relatives of broad totals, however, such as 'food'and 'clothing', understated the rise in the price of working-class purchases asrevealed by the 1937-8 budgets, since they referred to working-class purchasesbefore the first World War. Subsidies have been directed precisely to goodsheavily weighted in the old cost-of-living index, as it was much cheaper tostabilise this index than to stabilise the actual cost of working-class purchases.These price relatives would have, therefore, also been biassed (though in theopposite direction to the National Income White Paper's relatives), givingtoo low an estimate of Pr' and therefore too low an implied estimate of I.
The unbiassed estimates of I which remain indicate a value of I betweeni6i and i6z, using the methods described in Appendix B, a surprisinglylarge proportion of the estimates being at or very near this value. Examinationof the data indicates in fact that the information in the hands of the TechnicalCommittee led them to the coítclusion that the cost of buying the goods in the1937-8 working-class budgets was by June 1947 approximately 6i to 6zper cent higher than at the time these budgets were compiled.'
Now, having found I, we can use it to estimate the price relative for eachsection of the Index by modifying formula (2) thus
p, = - .1Wr
The full estimates for each section are shown in Appendix C, together with aderived set of estimates of how much separate items of the 1937-8 working-class budget would have cost in June 1947. A comparison between the newprice relatives for June 1947 inferred in this way, and the price relatives forsections of the old cost-of-living index for the same period, is made in this
'The Ministry of Labour seems open to criticism in continuing through the earlymonths of 1947 to publish an index showing an increase in the cost-of-living of approxi-mately 29 per cent for the same period, without stating that an increase over twice as greathad been derived by using a much more realistic weighting system.
THE INCREASE IN THE WORKING-CLASS COST-OF-LIVING 143
Appendix. It shows the combined effect of the distorted weighting system ofthat index together with the official policy of concentrating subsidies on itemswith a high weight in the index. The contrast is very marked for food,because of the underweighting of fruit, vegetables, etc. The old cost-of-livingindex only showed a price rise of 13 per cent, compared with the rise of38 per cent in the cost of the food in 1937-8 budgets. The marked rise in theprices of drink and tobacco was almost ignored by the old index.' Theconsequent bias in the old index has partially discredited official indexnumbers in the public's mind. This may eventually prove an expensive pricefor exploiting public confidence in the cost-of-living index, by selectingitems with high weights in the index for subsidising, so as to avoid thenecessity for a more rigorous price stabilisation policy.
The index number for each major group in the index is extracted andshown below. These index numbers can be linked by multiplication to thoseshown in the Ministry's Retail Price Index to bring them up to date, as isdone in the last column of the Table, which shows the index number at mid-March 1948, i.e. just before the Budget.
Working-Class Cost-of-Living, Pvc-War, June 1947 and March 1948
Although the e mphasis in this paper so far has been on inferring what theofficial estimate of the increase in the cost of living was, this estimate of 6x-6zper cent does have some intrinsic value. We should obtain an increase ofapproximately 6i-6z per cent if we used the weighting system described inthe Supplement (almost certainly the best available), the price relativesindicated in Table II below (from the old cost-of-living index for food andrelit items, and from the National Income White Paper for most of theothers), which would cover most of the Index's weights, and filled in theremaining unknown relatives with the help of the change of weights shownby the Supplement.3 Whatever was the Technical Committee's procedure,it would be difficult to select plausible weights and price relatives which
It is interesting, and somewhat sobering, for the statistician to note how the statisticaldecision before the first World War to under-represent tobacco in the old index and toexclude beer, affected fiscal policy and hence the living standards of the British public ageneration later.
* J is technically quite.correct to link the indices in this way, as the weights are the samethroughout. If q0 and p are the quantities bought and prices paid in 1937-8, and Piand p axe the prices of the same articles at June 1947 and any subsequent date,
Sqp0 q0p2 q,p1
q.p. lsP.O Professor R. G. D. Allen in fact estimates a 62 per cent rise from 1938 (Average) to
Mid-1947. London and Cambridge Economic Service Bulletin (Vol. XXV, Bulletin III, p.75).
193 7-8 Mid-June 1947 Mid-March 1948Food loo 138 150Rent loo 110 110Clothing loo 174 182Fuel and Light loo 144 158Household durables loo 206 222Misc, household goods loo 150 167Services loo 155 162Drink and tobacco loo 283 295
All iOO 161-2 172
¡44 THE BULLETIN
would indicate that the cost of buying the ¡937-8 budgets has not risen byabout this amount.
The estimates of the increased cost of the ¡937-8 budgets made by Mr.Nicholson and myself in previous issues of the BULLETIN' appear to havebeen not very inaccurate, though largely because of two opposite errors. Byusing the National Income White Paper indices, as can be seen from AppendixC, we over-estimated the price increases in food, clothing and fuel, but onthe other hand we considerably under-estimated the weights attributable todrink and tobacco, which meant giving the index a strong downward bias.On balance, my previous estimates of the war-time rise in the cost-of-livingwere too low.
A remarkable feature of the estimated increase in the weighted averagecost-of-living of 61-62 per cent from ¡937-8 to June ¡947, is that it is nearlyas much as the increase in the price of all consumers' goods, from the averageof 1938 to the average of 1947, given as 68 per cent in the latest NationalIncome White Paper. This is in striking contrast to the popular opinion thatprice policy, through subsidies on necessities and high purchase taxes onluxuries, has greatly assisted the working-class. The explanation seems to bethat before the war expenditure on tobacco was a larger proportion of totalexpenditure for the working-classes than for the whole population-5,0per cent as against 4.1 per cent; and so was expenditure on alcohol, 7.0per cent as against 4.3 per cent.2
The weighted average rise in the price indices for all items except drinkand tobacco was 53 per cent for the whole population, and 43 per cent forthe working-class. For all items, except drink and tobacco, the White Paperprice index was 1.07 times the working-class price index: for all items,including drink and tobacco, it was ¡.04 times. The large increases in theprices of drink and tobacco have therefore gone some way to offsetting theeffect of the discriminating price policy for other goods.
It must be emphasised that a cost-of-living index number, based on apre-war weighting system, has a limited meaning for two reasons. First,because of changes in quality, which cannot wholly be taken into account,but which were probably in the main downwards; and secondly, because theworking-classes are prevented by rationing from buying in ¡947 what theybought in ¡937-8, and have to buy substitutes which on the whole haveincreased more sharply in price, and which are often not entirely satisfactoryas substitutes.3 Both these qualifications cause what is in effect a downwardbias in the index, considered as an index of the cost of a defined living stand-
'BULLETIN, Vol. 7, No. 14, and Vol. 9, No. 7 respectively.Using the information on working-class drink and tobacco expenditure given in the
Supplement, and the distribution of total consumer expenditure shown in the NationalIncome White Paper.
Under ordinary circumstances such a wide dispersion of price relatives would cause aswitching of expenditure to goods whose prices had increased least, producing an upwardbias in the index. To remove the downward bias it will not be necessary for full freedom ofchoice to return, merely sufficient freedom for the same proportions of total expenditure tobe spent ou goods whose prices have risen less than average as in 1937-8 (i.e. for the corre-lation between price relatives and changes in weights to be zero).
THE INCREASE IN THE WORKING-CLASS COST-OF-LIVING 145
ard. This bias can be exaggerated as most of the items in the index are toostandardised for great changes in quality to have occurred; and working-classpurchases of foodstuffs, the chief rationed groups, were not always in1937-8 very much above the present rations, as is shown in Appendix D.It is calculated in this Appendix that if we replaced the 1937-8 weights forfood by weights corresponding to the rations in June iq,v we would obtainan increase of 67 per cent in the cost-of-living between 1937-8 and June'947.
Whatever the increase in the cost of buying the 1937-8 working-classliving standard between 1937-8 and June 1947 was, however, it was certainlymore than the 6,-6z per cent which we infer as the official estimate. If wemade a rough allowance for quality charges and restriction of choice, itwould seem to be reasonably accurate to say that the cost-of-living sodefined had risen between 1937-8 and June 1947 by at least two-thirds,and between 1937-8 and June 1948, by rather more than four-fifths.'
APPENDIX ATHE 1937-8 WEIGHTS
These can be determined with the aid of the information in the Supple-ment that the two sets of budgets of industrial and agricultural workers in1937-8 were combined in the proportion ,6: I (Supplement, p. 12); thatcertain items such as taxes, savings of different sorts, charities and bettingwere excluded (p. 4); that average alcohol and tobacco spending, because ofthe well-known tendency to underestimate these items in budgets, werere-estimated at ios. a week, divided in the proportions 53 per cent to beer,
per cent to spirits and wines, 33 per cent to cigarettes and 9 per cent totobacco (p. 6); that the weights appropriate to rice, sago, etc., coffee andanimal food could be considered spread proportionately over all food items;and that holiday spending, meals out, education, etc., could be distributedproportionately over all groups (the assumption being in each case that theirprice movements follow general price movements) (p. 4, footnotes).
With these clues as to official procedure, and the information on 193 7-8working-class budgets contained in the Ministry of Labour Gazette forDecember 1940, and January 1941, the expenditures and weights shown inTable I were calculated, with the corresponding weights for June 1947(from Appendix C. of the Supplement) shown for comparison. In some casesthe classification in 1947 is slightly different from that of the 1937-8 budgets,
1 The restriction on expenditure due to rationing does mean, however, that theseestimates must be used with some care. While they may not be very inaccurate as a measureof the cost of maintaining the 1937-8 standard, they clearly cannot be used for deflatingincomes which have increased by more than enough to maintain this standard, since anysuch excess must in fact be spent on unrationed goods whose prices have increased verysharply (notably drink and tobacco). Electricity seems at present to be the only majorcommodity which has become relatively very cheap, and whose consumption ic unrationed.
146 THE BULLETIN
and some rough minor adjustments in weights to meet these changes aredescribed in the footnotes to the Table. All cases in which weights are greaterin 1947 than in 1937-8 indicate that the price relatives used by the Ministryfor these commodities are greater than the estimated weighted average pricerise (and vice versa).
At first it seemed doubtful whether the Technical Committee hadrecommended the use of the clothing expenditure as shown by the quarterlyreturns of household expenditure or as shown by the special returns for awhole year of a smaller sample. When the weights were calculated on thefirst assumption, however, it was found that the weights for men's clothingwere higher than for women's (approximately 34 as against 29). Since thecorresponding weights for 1947 were higher for women than men (on anyreasonable assumption about the allocation of expenditure on clothingmaterials), this would indicate that women's clothing prices had increasedvery much more than men's had (about 30 per cent more, in fact). It seemsunlikely that any data on prices examined by the Ministry's statisticianswould lead them to use price relatives so different for men's and women'sclothing. The National Income White Paper indicates, for example, thatmen's clothing has risen in price slightly more than women's since pre-war.It can therefore be inferred with some confidence that the Technical Com-mittee used the special survey of each household's expenditure on clothingthroughout one year. The results of this survey are summarised in theissues of the Ministry of Labour Gazette that contain the summaries of thehousehold budgets.
(The decimal shown in the 1937-8 weight has little meaning itself, but inmost cases it will be the same as that which would have been reached by theMinistry, and allows subsequent calculations to be made rather more accur-ately).
FOOTNOTES TO PAGES 147 & 148.I Assumed divided equally between fish and potatoes for weighting.'as all in this section of manufactured foods because there was no specified
place in the budget for condiments, sauces, salt, jellies, custard powder, invalid foods, etc.' 2d. is also added to manufactured foods for soft drinks, for weighting.'The arrangement of 1947 weights is different from that of the 1937-8 budgets, and the
1947 weights are regrouped, the weight of 5 for' clothing materials' being allocated as 1 tomen's clothing, 2 to women's and 2 to children's.
'1d. is assumed transferred for weighting from gas to household appliances (hiring ofapparatus).
'id. is assumed transferred for weighting from electricity to household appliances(hiring of apparatus).
'3d. is assumed transferred from oil, firewood, etc., to miscellaneous household expendi-ture (matches).
'Miscellaneous expenditure is allocated as follows: 15d. to household appliances (sincethere is no place for radios, cycles, prams, clocks, electrical goods, etc.), 2d. to domestichardware, 2*d. to travel (petrol, etc.), 5d. to other services, and 5d. to be spread evenly(the last being required to fulfil the statement at the bottom of p. 6 of the Supplementthat about 3 per cent o, the weights are spread evenly).
$ Licences are believed over-estimated because one week fell in January. 2d. is assumedthe average weekly cost of radio licences (added to postage), and *d. allowed for motoringlicences (added to travel), other sorts of licences being in any case apparently excludedfrom the weighting system.
APPENDIX A
TABLE IWeights for Retail Price Index, 1937-8 and June 1947
Weight for
'47
Weight forGroup liera Average wee/dyexpenditure in 1937-8 (w) June 1947 (w)1937-8 (pence)
1 Bread 32.4 33.7 232 Flour 10.3 10.7 73 Cakes, etc
Biscuits14.13.9 32
4 Beef 32.0 33.3 255 Mutton and lamb 17.1 17.8 126 Pork 6.8Canned meat ......... 2.5
Sausages ......... 7.5 25.7 23Offal ............ 5.4Rabbits, etc.......... 2.4
7 Bacon, ham, etc 22.9 23.8 188 Fish,fresh ......... 7.1
dried ......... 2.0 11.6 8fried and chips' 4.1
9 Freshmilk .........Cream
36.0 3861.0 32
10 Butter 29.3 30.5 1911 Margarine 4.8)
Lard, etc 3.8 . 11.8 12Suet, etc. 2.7J
12 Cheese 8.4 8.8 513 SheU eggs 21.9 22.8 1314 Tea 19.7 20.5 1615 Sugar
Syrup, etc12.7 1431.0 11
16 Potatoes' 12.8 15.5 1117 Green vegetables ... 6.9)
Dried legumes 1.0 15.0 18Root vegetables 5.1 1Onions, etc 1.4
18 Apples 4.1Oranges 4.2Bananas ......... 2.7 19.0 23Other fresh fruits and nuts 3.1Dried fruits ......... 4.1
19 Oatmeal, etc.......... 2.5Fish, canned and paste 3.2Milk, skimmed 0.3
condwnsed 3.2dried ......... 15 35.6 4-0Cocoa ............ 1.5
Jam, marmalade, etc....... 6.2Canned and bottled veg 1.4Canned and bottled fruit 2.9Other food ' 9.4
- Rice, etc. 1.3 --Coffee 1.0 -- Food for animals 1.9 -
Total food" 393.5 407.7 348
II Rent 125.7 129.6 88
III Men's clothing.and materials 27.3 28.1 3j3Women's do 30.4 31.3 335Children's do 11.7 12.1 13Footwear 18.0 18.6 20
Total clothing5... 87.4 90.1 97
'Other items, assumed to follow the general price level:Meals, etc, out, 13.1; Education, music lessons, 3.6; Holidays. 7.4.Items excluded:Doctor, dentist, etc., 10.5; Hospital funds, 3.2; National insurance, 24.5; Insurance
premiums, 27.8; Trade union subscriptions, 15.7
148 TABLE I (continued)
Weighta for Retail Price Index, 1937-8 and lune 1947
Group Item Average wee/dy Weight for Weight forexpenditure in 1937-8 (w) June 1947 (w)1937-8 (pence)
IV 1 Coal 37.8Coke 1.0 4.0 4
2 Gas1 16.7 15.7 143 Electricity' 11.2 11.0 64 Oil, firewood, etc.' 9.2 6.4 4
Total fuel and light',',' 75.9 73.1 65
V 1 Furniture 12.7 13.1 182 Household utensils',',' 3.6 21.3 243 Floor coverings 6.8 7.0 114 Drapery 6.6 6.8 9S Ironmongery1 ...... 2.2)
Brushes and brooms 1.0. 7,4 9Pottery and glass 1.5 J
Total household durables',',' 34.4 55.6 71
VI 1 Soap (mcl. flakos) 9.1 9.4 82 Soda, cleansers, etc.' 4.2 7.4 93 Medicines 5.9 6.1 74 Newspapers
Books, stationery11.91 1482.5 f 11
Total miscellaneous goods' 33.6 37.7 35
VII 1 Rail faxes to work',' 7.1)Bus, tram fares to work 10.5 29.9 25Other fares 8.4J
2 Postage, etc.' ......... 4.9 7.1 93 Cinemas .........
Other entertainments10.35.5 16 3 18
4 Clothing repairs 1.7Boot and shoe repairs 6.8Hairdressing, etc....... 6.0 29.2 27Laundry ......... 6.2Domestic help' 2.6
Total services',' 70.0 82.5 79
VIII 1 Beer ... 63.6 'Spirits. etc. 6.0? 71.7 101
2 Cigarettes 39.6 STobacco 10.8)' 52.0 116
Total drink and tobacco 120.0 123.7 217
- Licences' 6.0-. Misc, expenditure' 30.0
GRAND TOTAL' 976.5 1,000 1,000
10*
APPENDIX '49APPENDIX B
PRICE RELATIVES AND IMPLIED Vues or I
The latest issue of the National Income White Paper (Cmd. 7371) showsa number of indices of price changes for different groups of consumer pur-chases. These indices are not entirely suitable because they refer to totalconsumer purchases, instead of working-class consumer purchases. And theyrefer to an increase from the annual aveiage prices during 1938 to the annualaverage prices during ¡947, whereas the relatives the Technical Committeewill have used are the increases from the average prices over certain datesfl 1937-8 (in October 1937, January 1938, April ¡938 and July 1938) to the
prices operating in mid-June 1947.The discrepancy in dates is probably the least important. The original
cost-of-living index (in those days somewhat less restrained in its movementsthan it became later) showed an average of 102 for the months indicated in1937-8, as against ioi for the average of 1938.' The pace of price rises seemedto accelerate in the autumn of 1947 because of the combined effects of theautumn budget's increase in purchase taxes and the ceiling imposed oncost-of-living subsidies. So possibly prices in June 1947 were slightly belowthe average for the year. The wholesale prices index, for example, was at187.2 in June 1947, compared with an annual average of 189.1 Tobaccoprices would be an exception, and the relative for them has to be estimatedseparately. The effect of the date discrepancies, therefore, might be to causethe White Paper price relatives and therefore the implied estimates of I, thecost-of-living index derived from them, to exaggerate increases but only veryslightly. These price relatives and the values of I they imply, using formula(z), are shown in section A of Table II below. In order to allow an assess-ment of the difference in scope between the groups in the White Paper andin the Retail Price Index, notes on coverage are also given in Table 11.2
A source of error lies in the fact that the weights are in some cases toosmall for I to be estimated accurately, since W is given only in wholenumbers. The extreme case is that of electricity, in which the range ofpossible values of lis 147.2 to 174.0 (Wr having a possible range of 5.5 to 6.).The weights in June 1947 are therefore also shown in the Table.
The price rises shown by the sections of the original cost-of-living indexfor food and a few other commodities are given in section B of the Table.The index number for June i7th, 1947, was compared with the average ofthe index numbers of November ist, 1937, February ist, 1938, April 3oth,1938, and August znd, I938these being the nearest dates of the weeks forwhich the working-class expenditure budgets were provided. A column ofNotes indicates diffetences of coverage, and the last column, as for section A,shows the valueof I obtained by using the price relative concerned in formula(z).
'Annual Abstract of Statistics, No, 84, Table 302.2 The notes on the Retail Price Index group refer generally to the goods the Committee
arc usiag for calculating current price relatives. 1 resumably however the price relative forJ une 1947, used for changing the weighting system, would have been chosen where possibleso as to be comparable to the cuiTent ones.
Gro
up
(1)
Food
Ren
t and
Rat
es
TA
BL
E I
IE
stim
ates
of
the
impl
ied
incr
ease
in th
e co
st-o
f-liv
ing
(I)
A.
Usi
ng N
atio
nal I
ncom
e W
hite
Pap
er g
roup
indi
ces
Whi
te P
aper
Gro
up N
umbe
rIn
der
of R
etai
l Pri
ces
Gro
upan
d N
otes
Num
ber
and
Not
es
(2)
1. V
ery
hete
roge
neou
s, a
nd s
ub-
ject
to v
ery
diff
eren
tial p
rice
polic
y.4.
'Inc
lude
s an
allo
wan
ce f
orbu
ildin
gs o
ccup
ied
by n
on-p
rofi
tm
akin
g bo
dies
and
for
hot
els.
boar
ding
hou
ses,
etc
.'B
ased
on S
ched
ule
A a
sses
smen
ts, i
n-cl
udin
g re
ntal
val
ue o
f ow
ner-
occu
pied
hou
ses.
Men
's8b
(i)
Bot
h m
en's
and
boy
'sC
loth
ing
clot
hing
.V
ery
hete
roge
neou
s.Su
bjec
tto
diff
eren
tial
pric
epo
licy
(an
are
all c
loth
ing
grou
ps.)
Wom
en's
Sb(i
i)V
ery
hete
roge
neou
s.C
loth
ing
Cov
ers
girl
s' a
nd in
fant
s' c
loth
-in
g (i
nclu
ding
dre
ss m
ater
ials
,kn
ittin
g w
ools
and
hab
erda
sh-
ery)
, as
wel
l as
wom
en's
clo
thin
g.A
ll C
loth
ing
8b.
Foot
wea
r8e
. Het
erog
eneo
us.
AU
CL
othi
ng8.
Het
erog
eneo
us.
and
Foot
wea
rC
oal
5e. A
ll do
mes
tic c
oal,
incl
udin
gm
iner
s' c
oal a
t pith
ead
pric
es.
Gas
5e.
Ele
ctri
city
5b.
Furn
iture
and
6e.
Ver
y he
tero
gene
ous.
In.
Furn
ishi
ngs
dude
s w
irel
ess
sets
and
mus
ical
inst
rum
ents
.
(3)
I. M
uch
mor
e itr
ictly
con
fInr
dto
wor
king
-cla
ss p
urch
ases
.
II.
A v
ery
high
prop
ortio
nw
ould
be
cont
rolle
d re
nts,
and
owne
r-oc
cupi
ed d
wel
ling
rent
sre
lativ
ely
few
.
1H. 1
,2,
and
one
-fif
th o
f 7.
Men
's c
loth
ing
and
mat
eria
ls.
sele
ctin
g 'th
e lin
e w
hich
is p
re-
dom
inan
tly s
old
to w
orki
ng-
clas
s ho
useh
olds
'.II
I. 3
, 4, a
nd tw
o-fi
fths
of
7.W
omen
's c
loth
ing,
and
clo
thin
gm
ater
ials
onl
y. S
ame
qual
ific
a-tio
n re
nar
row
ness
of
scop
e as
for
mal
e cl
othi
ng.
III.
1-7
.II
I.8.
9 a
nd 1
0.Si
mila
rly
rest
rict
ed in
sco
pe.
III.
Res
tric
ted.
IV. 1
. Inc
lude
s co
ke.
IV. 2
.IV
. 3.
V. 1
, 3 a
nd 4
.
June
194
7W
eigh
t (w
)w
Whi
tePa
per
pric
e re
la-
tives
(p)
Impl
ied
valu
e of
I
(4)
(5)
(6)
(7)
348
1.17
153
179
881.
4710
815
9
310.
9119
417
6
330.
9519
218
2
770.
9319
217
820
0.93
175
163
970.
9318
917
5
410.
9816
616
2
141.
1214
516
36
1.83
8716
038
0.71
231
163
Dom
estic
6b.
Ver
y he
tero
gene
ous.
Ex-
hard
war
edu
des
wir
eles
s se
ts a
nd m
usic
alin
stru
men
ts, a
ndcy
cles
and
spor
ts g
oods
.So
ap.
7b. I
nclu
des
scou
rers
.H
ouse
hold
7.M
atch
es,
soap
,sc
oure
rs,
non-
dura
bles
, pol
ishe
s, c
andl
es, c
lean
ing
mat
er-
ials
, etc
.R
eadi
ng9.
mat
ter.
Tra
vel.
11. I
nclu
des
taxi
s.
Ent
erta
in-
13. A
dmis
sion
cos
ts to
all
ente
r-m
ents
.ta
inm
ents
, cin
emas
bei
ng th
em
ajor
one
s,M
isc.
14. M
edic
al (
othe
r th
an N
.H.I
.),
serv
ices
,un
dert
akin
g, h
otel
s, h
aird
ress
-in
g, d
omes
tic, a
ll re
pair
s, a
ssur
-an
ce c
osts
, pro
pert
y tr
ansf
ers,
hire
of
appl
ianc
es,
etc.
Ex-
trem
ely
hete
roge
neou
s.A
lcoh
ol2.
Com
preh
ensi
ve.
(Usi
ng 1
937-
8 B
udge
t wei
ghts
for
beer
and
'oth
er' t
o re
cai-
cula
te c
ombi
ned
Whi
te P
aper
Inde
x).
(With
pri
ces
calc
ulat
ed o
n th
eba
sis
of v
olum
e of
liqu
id in
stea
dof
alc
ohol
con
tent
.)*
Tob
acco
3.C
ompr
ehen
sive
(in
clud
ing
gift
s to
troo
ps o
vers
eas)
.(U
sing
193
7-8
Bud
get w
eigh
ts f
ola
te c
ombi
ned
Whi
te P
aper
iode
(Adj
ustin
g fo
r in
crea
sed
pric
e in
V. 2
. and
5.
Incl
udes
wir
eles
sse
ts a
nd m
usic
al in
stru
men
ts,
and
also
cyc
les
and
spor
ts g
oods
.(T
he 1
937-
8 w
eigh
t is
doub
tful
.)V
I. 1
. Soa
p on
ly.
VI.
1an
d 2.
Mat
ches
, soa
p,sc
oure
rs a
nd p
olis
hes.
VI.
4. I
nclu
des
6tat
ione
ry.
VII
.1.
Incl
udes
mot
orin
glic
ence
s(a
ndpo
ssib
lyso
me
mot
orin
g ex
pens
es).
VII
. 3. P
redo
min
antly
cin
ema.
VII
. 4.
Rep
airs
to f
ootw
ear,
laun
dry,
hai
rdre
ssin
g, d
omes
tic,
etc.
(19
37-8
wei
ght d
oubt
ful)
.
VII
I. 1
.B
eer
rela
tivel
y m
uch
mor
e im
port
ant.
VII
I. 2
.Pr
esum
ably
exc
lude
sex
pens
ive
bran
ds.
r ' C
igar
ette
' and
oth
er, t
o re
calc
u-x)
. mid
-194
7).f
330.
8720
017
4
81.
1813
716
117
0.99
150
148
111.
3512
016
1
251.
2012
715
2
180.
9117
616
0
271.
0814
916
1'e z X
101
0.71
250
178
(253
)(1
80)
(200
)(1
42)
116
0.45
333
149
(336
)(1
51)
(376
)(1
69)
* T
he p
revi
ous
Whi
te P
aper
(C
md.
709
9) s
how
ed a
rel
ativ
e of
258
for
194
6, w
hich
was
ass
ocia
ted
with
a r
elat
ive
of 2
09 f
or a
give
n vo
lum
e of
bee
r (i
.e. f
or a
'bul
k' b
arre
l, in
stea
d of
a s
tand
ard
barr
el).
Sin
ce p
rice
s di
d no
t gre
atly
cha
nge
in 1
947,
and
othe
r al
coho
l pri
ces
had
a lo
wer
rel
ativ
e, it
can
be
estim
ated
that
on
a vo
lum
e ba
sis
the
rela
tive
for
1947
mig
ht h
ave
been
abo
utth
e 20
0 sh
own
for
all a
lcoh
ol.
It w
as a
ssum
ed th
at o
ne-t
hird
of
1947
toba
cco
cons
umpt
ion
was
con
sum
ed a
t 194
6 pr
ices
(al
low
ing
for
a tim
e la
g be
fore
toba
cco
pric
e ri
ses
wer
e ef
fect
ive,
and
for
the
fall
in c
onsu
mpt
ion)
, whi
ch im
plie
s th
e pr
ice
inde
x sh
own
for
June
194
7.B
ecau
se o
f th
e el
astic
dem
and
for
eggs
. The
fir
st in
dex,
like
the
othe
rs s
how
n, is
bas
ed o
n th
e av
erag
e of
the
four
pri
ces
atth
e da
tes
indi
cate
d ab
ove.
TA
BL
E I
I (c
ontin
ued)
B.
Usi
ng C
ost-
of-l
ivin
g In
dex
Pric
e C
hang
es.
(Jul
y 19
14=
100)
Com
mod
ityN
otes
on
scop
e of
cos
t-of
-liv
ing
inde
x.In
dex
Inde
xPr
ice
June
194
7Im
plie
d(R
etai
l Pri
ce19
37-8
June
Rel
ativ
eW
eigh
tw
valu
e of
IIn
dex
Gro
up19
47(p
)(w
)(1
)(2
)(3
)(4
)(5
)(6
)N
umbe
r)I.
1. B
read
161
178
111
231.
4716
24.
Bee
f(M
ean
of r
ib a
nd f
lank
indi
ces)
132
1521
115
251.
3315
55.
Mut
ton
(Mea
n of
Bri
tish
and
Fore
ign
legs
and
and
lam
bbr
east
indi
ces)
133
141
105
121.
4815
5
9. F
resh
milk
195
254
13'
321.
2115
710
. But
ter
(Mea
n fo
r fr
esh
and
salt)
1151
116}
101
191.
6116
213
. Egg
s15
413
909
013
1.75
158
(Usi
ng p
rice
per
egg
sho
wn
by th
e bu
dget
s)4
(125
)(1
11)
(195
)14
. Tea
148
185
125
161.
2816
015
.Sug
aran
dSu
gar
only
.sy
rup
1241
145
116
111.
3015
1
16. P
otat
oes
136
171
126
111.
4117
87.
Bac
on a
ndB
acon
onl
yha
m13
8116
712
118
1.32
160
8. F
ish
Het
erog
eneo
us c
omm
odity
2051
238
116
81.
4516
812
. Che
ese
126
116
092
51.
7616
2T
otal
Foo
dV
ery
hete
roge
neou
s, it
ems
in th
e ol
d in
dex
bein
g sp
ecia
lly li
able
to s
ubsi
dy.
142
161
113
348
1.17
132
II. R
ent&
Rat
es E
xclu
des
hous
es b
uilt
sinc
e 19
14.
159
175
110
881.
4716
21H
. Clo
thin
gV
ery
rest
rict
ive,
exc
ludi
ng w
omen
's o
uter
-w
ear,
chi
ldre
n's
clot
hes,
slip
pers
, etc
.21
034
516
497
0.93
152
IV. 1
. Coa
l, C
oke
194
307
158
410.
9815
42.
Gas
149
217
146
141.
1216
4IV
. Tot
al f
uel a
nd li
ght
(exc
i. el
ectr
icity
)18
028
315
759
1.05
165
The mean, median and mode of this distribution are respectively ¡62.7,i6z and i6,.8. That all values of I are not identical is due to random errorsin rounding and biassed errors in the price relatives used.' Since the medianand the mean depend on whether more relatives with an upward bias thanwith a downward bias are used, the mode is the most appropriate measure.However it seems preferable to proceed by eliminating price relatives whichwe can expect to be biassed.
For alcohol the Supplement states that'It is possible, however, to adjust the price each month to allow for
the extra duty which would have been payable if it (beer) had been of thestrength which prevailed in ¡938, and this adjustment goes some waytowards giving a series reflecting the movements in the price of beer ofunchanged quality' (p. 26).
The resultant series is described as a ' compromise ' between the alternativesof a series of beer prices ignoring changes in strength (i.e. on a 'bulk barrel'basis), and a series of the price of a certain volume of alcohol (i.e. on a'standard barrel' basis). The series used 'may be said to reflect the factthat the purchaser of a glass of beer is not only concerned with the alcoholiccontent, but also with the quantity of liquid refreshment, service, recreation,social intercourse and amenities of the public house.' The Technical Com-mittee would presumably have used a price relative constructed in the sameway for calculating the ¡947 weight, so that the implied value for I would besomewhere between the values of i and i8o shown in Table IIvalueswhich would have been implied by the change in weights if respectivelyconstant volume and constant alcohol content had been used as bases forcalculating the price relatives. For tobacco a precise calculation cannot bemade. If only standard brands of cigarettes are considered, the price increaseis from 6d. to ¡s. 8d. for ¡o, which gives a price relative of 367 (comparedwith 365 for' all cigarettes' by the method indicated in the footnote to TableII) implying a value for I of ¡64, but the inclusion of sub-standard brands
1 Bias being due to the fact that the price increase in working-class purchases is notproperly measured either by the increased price of all consumer purchases, or by the increasein the prices of goods represented in the old cost-of-living index.
'The drafter appears to take the slightly more idealistic view on the functions of thepublic house.
APPENDIX B
If we group the implied estimates of I, we have
¡53
Implied Estimate Number of Implied Estimate Number ofof I Estimates of I Estimates130-4 ... I 170-4 1
135-9 ... - '75-9 ... 6140-4 ... I ¡80-4 2¡45-9 2 185-9 ... -¡50-4 ... 5 190-4 ... -'55-9 ... 5 '95-9 ... I-¡60-4 ... ¡6i6-ç 3
Total 43
154 THE BULLETIN
and various tobacco brands might indicate a different relative, and the methodused cannot be inferred. So the values implied by alcohol and tobacco for Imust be eliminated.
In section A we can eliminate food, clothing and household durables asbeing groups which are heterogeneous and subject to a differential pricepolicy. The Technical Committee would use other and lower indices forworking-class purchases than those given in the White Paper. The groups offurniture and domestic hardware are too different from those of the indexfor the price relatives to be necessarily applicable to the goods covered bythe index sections.
The section A groups which are left are presumed to give values of Imainly free from bias, as far as differences of scope are concerned. Therandom errors due to the use of whole numbers for the second set of weightsare still present, but to deal with these we indicate, as well as the impliedvalue, the range of implied values of I which are possible.'
Of these ten groups, seven could be satisfied by any value of I in the range158.2-163.9 (though not the same 7 by all values in that range since in therange i58.z to 159.9 rent would be satisfied but not coal). It seems obviousthat I cannot be below 155, and the implied values given by household non-durables and travel can therefore be ignored. (They are categories whichcontain differences of definition that make them of doubtful value anyway).The mean of the implied values of I for the remaining categories is j6o.8.
Bias in the groups of section B, though this time in a downward direction,can be expected in the same groups as for A, namely total food and clothing,because of the tendency to direct subsidies towards items included in the oldcost-of-living index. The results from meat and butter groups have to bediscounted because of the difficulty of knowing what weights would be givento the different types,2 and total 'fuel and light' is not comparable with theindex group. This leaves the following implied values for 1:3-
Eight of these twelve groups would be compatible with a range of I from159.2 to 159.3 or i6i.o to 162.9, and seven by 158.3 to 159.2, 159.3 to i6i.oor 162.9 to 164.2 This increases the presumption that I is within the range
1 By using the highest and lowest possible values of W in formula (2).2 The price relatives for foreign mutton for example were 93 for breast and 115 for legs.'The Technical Committee appears, incorrectly, to have ignored the elasticity of demand
for eggs, since a price relative allowing for elasticity gives a quite extreme value for T.
Impliedvalue of I
Range ofimplied values
Impliedvalue off
Range ofimplied v4lues
Bread ... 162 ... 159.2-166.3 Bacon 160 ... 155.6---164.6Milk ... 157 . . 154.4-159.3 Fish ...... 169 ... 158.3-179.5Eggs ... 158 ... 152.0-164.2 Cheese ...... 162 ... 147.2.180.0Tea ... 160 ... 155.3-165.4 Rent and Rates 162 ... 161.0-162.9Sugar ... 151 ... 144.2-158.0 Coal ...... 154 ... 152.5-156.3Potatoes 178 ... 169.8-186.0 Gas ...... 164 ... 158.1-169.8
Implied Range of Impliedvalue of I implied values value of I
Range ofiml,lied values
Rent and rates 159 ... 158.1-159.9 Household non-durables 148 ... 144.8-152.7Coal ... 162 ... 160.0-164.0 Reading matter 161 ... 154.4-169.2Gas ... 163 ... 157.0-168.6 Travel 152 ... 148.9-155.0Electricity 160 ... 147.2-174.0 Entertainments 160 ... 155.1-163.9Soap ... 161 ... 151.5-171.7 Miscellaneous services 161 ... 158.2-164.2
APPENDIX B 155
158.2 to 163.9 indicated by the unbiassed values in section A, and suggeststhat the cost-of-living price relative was not used for coal or sugar (the newweighting system in any case referring to ' coal and coke' and ' sugar andsyrup '). The high value of I implied by the potato index suggests that itwas mistaken to allocate' Fish and Chips ' equally between fish and potatoes.If potatoes were allotted id. instead of 2d., for example, an implied valuefor I of 164 would be possible.' This however also means that the value ofI implied by the price relative for fish would have to be adjusted upwards, sothat the value shown must be discarded. For the remaining groups, the meanof the implied values of Jis 160.6.
It seems therefore that the Technical Committee probably compiledindividual food price relatives in most cases from the data used to computethe cost-of-living index, and fuel and light price relatives from the materialused in reaching the National Income White Paper estimates. It would alsoappear more likely that they used the rent price relative implied by the cost-of-living index, than by the National Income White Paper, and price relativessimilar to those used in the National Income White Paper for soap, entertain-ments and miscellaneous services.
These give us, from both sections, 14 estimates of I ranging from 157to 163. The estimates implied by milk (i') and eggs (158) are, however,the only ones under 159, and it seems that a different price relative has beenused from the ones obtained above. (These commodities are particularilysubject to seasonal influences and it may be that the cost-of-living priceindices have been used, but with different base dates). This leaves 12 esti-mates of I ranging from i6o to 163 and with a mean of 161.2. Thevalues of i6i and 162 would lie in the possible range of all 12. These valueswould also satisfy other relatives which have been excluded because ofdifficulties of definitionfootwear and furniture from section A, and butterfrom section B, and they are consistent with a relative for alcoholwhich is half-way between those given by 'volume' and 'alcohol content' bases.
The value of I given by equation (2), using whole number values for w,is not in fact the best estimate of I. Since lis an inverse variable of w, andw is equally likely to have any value over a range 0.5 each side of the wholenumber shown, the expected value of I is not the value given by the centreof that range (i.e. the whole number. If!' is the expected value of I, we have,from equation (z)
Wr + 0.5dw
JwW-0.5I =wrpr =
ZUr + 0.5
J
Wr 0.5The mean I' from all the eleven groups is 161.3.2
Possibly some allowance was made for cooking fats used in frying fish and chips.*i assumes that price relatives for these groups are equally likely to approximate to
thne fh Tp,hnfr,.I Cí,mmift.p
W +0.5WrPr loge
Wr 0.5 (4)
156 THE BULLETIN
By different methods, therefore, we obtain estimates for I of i6i.8, 161.2and 161.3. So it seems highly probable that in the Technical Committee'sopinion the cost-of-living in June 1947 was approximately 6, to 6z per centabove that at the time the budgets were collected in 1937-8. It is extremelyunlikely, for example, that their conclusion was that the cost-of-living was,say, only 55 per cent higher, since this conclusion would have involved usingdifferent price relatives for a number of homogeneous commodities (such asbread, tea, bacon and fuel) than those most suitable and most readily avail-able. Such a value would imply an increase in 'rent and rates' of only about5 per cent, even lower than the 10 per cent shown by the cost-of-living index.It would moreover hardly have been compatible with the price relatives forother homogeneous commodities, since the approximation involved in usingwhole number values for w would be most unlikely to cause persistenterrors in estimating Iall in the same direction.
APPENDIX CINCREASES IN PRICES OF COMMODITIES BOUGHT BY THE WORKING-CLASS
Table III shows the price relatives for each commodity group in theindex obtained by using formula (i), expressed as index numbers.' In somecases, where the weight attributable to each item in 1937-8 is not knownprecisely, because of regrouping, the index is not firm, and is thereforeshown in bracketsz_though the amount of regrouping between totals ofgroups was not sufficient to affect their relatives greatly, and these indexnumbers are therefore shown without brackets. For comparison, the secondand third columns show roughly corresponding index numbers for each itemfrom the old cost-of-living index3 and the National Income White Paperrespectively.
The right-hand side of the Table shows first the amount spent by theaverage working-class household on each commodity in 1937-8, and then, inthe last column, the Cost of the same goods in June 1947. This colunm doesnot depend on the grouping of the items of the 1937-8 budget, since it can bederived directly from the June 1947 weight, which is published, and the totalcost of the 1937-8 budget at June 1947, which can be found from I and theoriginal cost of the 1937-8 budget. The figures in the last column must bethe approximate base expenditures for the new Retail Prices Index.
The dispersion of price relatives used indicates how carefully theTechnical Committee estimated price changes for calculating the newweights, and nence that their estimate of I is much more valuable than theSupplement alleges.
1 These are estimates of the price relatives used by the Technical Committee, which maynot be very well based in all cases, though the information at the Cominitte's disposal is ofcourse, considerable.
'In addition, there may be random errors in individual items due to the rounding, ofthe June 1947 weights. especially where these weights axe small.
'Calculated as in Table II.
TA
BLE
III
Pric
e In
dice
s fo
r W
orki
ng-C
lass
Pur
chas
es, a
nd th
e co
st o
f the
ave
rage
hou
seho
ld b
udge
t of 1
937-
8, a
t 193
7-8
and
at J
une
1947
tric
es.
(Tot
als
do n
ot a
lway
s eq
ual t
he s
um o
f ite
ms,
bec
ause
of
roun
ding
).
Est
imat
ed w
ork-
ing-
clas
s pr
ice
inde
x fo
r w
id-
twne
194
7
-8=
Old
cos
t-of
-liv
ing
Nat
iona
l In-
Ave
rage
inde
x fo
r m
id-
com
e W
hite
wee
kly
June
194
7Pa
per
pric
eex
pend
iture
inde
x fo
r av
e.19
37-8
of 1
947
(193
7-8=
1001
(193
8 av
e. 1
001
Cos
t of
sam
e go
ods
June
194
7
110
106
111
s.d.
28 10
s.d.
30
1127
516
4112
1*
2833
109
t15
1714
521
3012
212
1'11
124
(111
)11
6(
11)
10
134
130
31
42
101
2525
164
1117
9292
88
9290
110
19
126
125
18
21
124
116e
12
15
(115
)12
6(1
3)1
519
412
2419
51
63
0(1
81)
(32)
52
(138
)(
4)6
138
113
153
330
455
110
110
108
106
116
(178
)19
4g2
3(4
1)(1
70)
192'
26
(44)
(173
)1
0(1
8)
174
192
59
10 0
174
175
16
27
174
164
189
73
128
* R
ibs
109,
fla
nk 1
23.
f B
rit.:
legs
109
. bre
ast 1
01. F
orei
gn: l
egs
115,
bre
ast 9
3.Fr
esh
93, s
alt 1
09.
IFo
od 1
Bre
ad2
Flou
r3
Bis
cuits
, cak
es, b
uns,
etc
.4
Bee
f5
Mut
ton
and
lam
b6
Oth
er m
eat..
....
7B
acon
and
ham
8Fi
sh9
Fres
h m
ilk ..
....
10Il
utte
r11
Oth
er f
ats
12C
hees
e13
Egg
s14
Tea
15Su
gar
and
syru
p16
Pota
toes
17O
ther
fre
sh v
eget
able
s18
Fres
h an
d dr
ied
frui
t19
Mis
c. m
anuf
. foo
dsO
ther
foo
ds
Tot
al F
ood
Ren
t and
Rat
es
Clo
thin
g:M
en's
clo
thin
g an
d m
ater
ials
Wom
en's
do.
...C
hild
ren'
s do
.
All
Clo
thin
gFo
otw
ear
Tot
al C
loth
ing
and
Foot
wea
r ...
...
TA
BL
E I
II (
cont
inue
d)Pr
ies
Indi
ces
for
Wor
king
-Cla
ss P
urch
ases
, end
the
cost
of
the
over
age
hous
ehol
d bu
dges
of
1937
-8, a
t 193
7-8
and
aS J
une
1947
pric
es.
(Tot
als
do n
ot a
lway
s eq
ual t
he s
um o
f ite
ms,
bec
ause
of
roun
ding
).
IV F
uel a
nd L
ight
1C
oal a
nd c
oke
2G
as3
Ele
ctri
city
4O
il, f
irew
ood,
etc
.
Tot
al F
uel a
nd L
ight
V H
ouse
hold
dur
able
s:1
Furn
iture
2A
pplia
nces
3Fl
oor
cove
ring
s4
Dra
pery
5 H
ardw
are,
pot
tery
, gla
ss
Tot
al H
ouse
hold
Dur
able
s
VI
Mis
cella
neou
s G
oods
:1
Soap
.....
......
.2
Soda
, cle
anse
rs, m
atch
es, e
tc.
3M
edic
ines
, toi
let g
oods
4 R
eadi
ng m
atte
r
Tot
al M
isce
llane
ous
Goo
ds
VII
Ser
vice
s:1
Tra
vel
2Po
stag
e, e
tc.
3E
nter
tain
men
t4
Oth
er s
ervi
ces
Tot
al S
ervi
ces
Est
imat
ed w
ork-
Old
cos
t of
livin
gN
atio
nal I
n-A
vera
geC
ost o
fin
g-da
ss p
rice
inde
x fo
r m
id-
com
e W
hite
Wee
kly
sam
e go
ods
inde
x fo
r m
id-
June
194
7Pa
per
pric
eex
pend
iture
June
194
7Ju
ne 1
947
inde
x fo
r av
e.19
37-8
of 1
947
Av'
,,,
'"I.
166
158'
166'
33
5 4
(144
)14
614
5(1
3)1
10(8
8)87
(11
)9
(101
)(
6)6
144
157"
157
311
8 6
222
231'
11
24
(182
)(1
9)3
125
47
1521
47
12(1
96)
200'
(7)
12
206
218
46
93
137
137'
91
1(1
96)
(7)
12
185
611
120
1201
01
21
5
130
31
47
(135
)12
7*1
(26)
33
(202
)(
7)1
217
817
61
42
4(1
49)
149
(24)
36
135
68
104
1B
acon
onl
y.I C
alcu
late
d on
an
aver
age
quar
terl
y pr
ice
for
1937
-8 (
not t
he a
vera
ge p
rice
pai
d).
Suga
r on
ly.
'Cof
fee,
sag
o, r
ice,
etc
., an
d fo
od f
or a
nim
als,
exc
lude
d fr
om th
e w
eigh
ting
syst
em f
or f
ood.
Men
's a
nd b
oys'
wea
r.W
omen
's, g
irls
' and
infa
nts'
wea
r.'C
oal o
nly.
"Exc
I. e
lect
rici
ty.
'Fur
nitu
re a
nd f
urni
shin
gs.
IIn
clud
es a
pplia
nces
.'In
clud
es s
cour
ers.
10E
xclu
des
stat
ione
ry.
ItE
xclu
des
mot
orin
g (e
.g. m
otor
ing
licen
ces)
.jt
On
an a
lcoh
olic
con
tent
bas
is (
the
rela
tive
in th
e fi
rst c
olum
n ap
pare
ntly
take
s vo
lum
e of
liqu
id a
lso
Into
acc
ount
).E
duca
tion,
mus
ic le
sson
s, h
olid
ay e
xpen
ditu
re, o
ther
lice
nces
and
mea
ls o
ut, a
ll of
whi
ch a
re a
ssum
ed b
y th
e T
echn
ical
Com
mitt
ee to
fol
low
the
gene
ral i
ndex
and
are
exc
lude
d fr
om th
e w
eigh
ting
syst
em.
V'
VII
I D
rink
and
Tob
acco
1A
lcoh
olic
dri
nk2
Tob
acco
Tot
al D
rink
and
Tob
acco
Oth
er e
xpen
ditu
re13
GR
AN
D T
OT
AL
227
360
2501
*33
3
s.d.
510
4 2
s.d.
132
152
283
286
10 0
284
(161
-2)
2945
161-
212
916
883
713
411
i6o THE BULLETIN
APPENDIX D
WORKING-CLASS FOOD PURCHASES, PRE-WAR AND IN JUNE 1947
The average persons per household shown by the 1937-8 budgets (assum-ing that one-third of those under ¡4 were under 5, and weighting industrialand agricultural households in the ratio i 6 to i) were as follows :-
Weighting 1937-8 food purchases of industrial and agricultural householdsalso in the ratio i6 to ¡, and multiplying the rations available on differentration books in June ¡947 by the number of persons in each age group, we candraw up Table IV, which compares weekly purchases of a standard householdbefore and after the war. (Note that the ¡937-8 budgets were only collectedfrom the households of those in employment. Average working-class pur-chasers, including the unemployed, would have been lower than thoseshown here.
TABLE IVWeekly purchases of sorne ,nain foodstuffs by a Standard Household of 3.77 Persons,
1937-8 and June 1947-(lbs.)
On the assumption that there is one adult manual worker in the household, its totalB.U.s per week would be about 39. These are assumed to be spent so that the quantitiesof bread, flour and cakes (mcl. buns) were in the same proportions as were shown by theaverage households in 1946 (vide' \Vorking-Class Income and 1-lousehold Expenditure ', byMiss T. Schulz, BULLETIN, Vol. 9, PP. 133 et seq.). This would involve an outlay of B.1.'.sas follows: Bread, 32 Flour, 4; Cakes and Buns, 2F
2 A price of id. per ounce of meat is assumed, Bacon only.Excluding special priorities, school milk, etc.
'Margarine could be taken instead of butter, and margarine consumption would berather higher than butter.
'Excluding special allocations.Based on a ration of 61 per annum for those over 2.
Average purchases193 7-8
Rationsjune 1947
Bread 13.6 13.9'Flour 4.5 1.4'Meat 4.7 about 3.6'Bacon, ham 1.4 0.5'Milk (pints) 11.0 13.1Butter 1.8 0.8'Margarine 0.7 0.8'Lard 0.5 0.3Cheese (sold by weight) 0.7 0.5Tea 0.7 0.5Sugar 4.9 1.9'Jam and marmalade 1.0 0.9Shell eggs (no.) 13.8 about 477
Age 0- 5 0.33-i8 0.98
2.46
Total 3.77
APPENDIX D iói
The effect of food restrictions can he roughly indicated by estimating howmuch the cost-of-living would have risen if the working-class household hadonly b ght these rations at both dates. Strictly we ought to allow for otherrationing also, but it is doubtful whether working-class clothing and fuelpurchases are greatly reduced below pre-war standards by rationing, and inany case fuel and clothing do not differ greatly from the average price rise-the distorting effect of food restrictions being due to the fact that they apply togoods whose price has increased least. If we reduce the weighting of the fooditems shown in Table IV, by the proportions indicated, increasing the weight-ing of other items correspondingly, and use the price relatives given in TableIII, we get an index number of 167.0. This figure measures the increasedcost of the 1937-8 living standard on the assumptions that money saved byreducing spending in 1937-8 (if expenditure in food had been restricted thenby rationing like that in force in June, 1947) could have been spread propor-tionately over all other purchases without net loss of satisfaction, and thatexpenditure coild be distributed in this way in June '947. This is, ofcourse not likely to be true, since it is presumed that the actual budget in1937-8 represented a maximisation of satisfaction (so that a shift of spendingin any way must therefore of itself imply a loss of satisfaction). Also theweighting system so adjusted still did not represent a possible distributionof expenditure in June 1947, because of other shortages (e.g. food on points,clothing and fuel). It does however indicate the magnitude of the error intro-duced by rationing.'
DUDLEY SEERS.
'It seems that meat, bacon, fats, sugar and eggs were the only foodstuffs for whichworking-class consumption was considerably reduced below pre-war levels by rationing: