HOUSE RENTS

Evening Post, Volume CVII, Issue 82, 5 April 1924, Page 13

HOUSE RENTS

STATISTICIAN'S COM-

PILATION

VALIDITY'.■'OF THE INDEX

AN EXPLANATORY- STATEMENT.

1 Interest in tie question- of house | rents'is again, likely to be aroused by i the publication of the'results of the I ■ half-yearly rent collection for. Feb- . I ' ruary,- 1924..- These figures are pubI jlished in the March issue, of'the } '"Abstract of Statistics,'.' and to I , clear away any misconceptions exI \isting. as .to the. validity of'the rent | -index,' the Government. Statistician ■ (Mr. Malcolm Fraser)-lias issued an \ explanatory statement-, -showing how I ' the comparison between the rent i. . statistics and . the Census rents,. j which canse'd such a. controversy in j/. ■ December1 last; has been arrived at. i. , Statistics of retail, prices, or what is'

j commonly referred to as "cost of living,", j were first.'initiated by the Census and Statistics) Office in 1914. The system \ followed is that known as the "aggrei gate-expenditure method," and Was fully I explained in the first report issued on the subject of priocsi; The inquiry, first coy-. ' pred food, rent, and fuel, and light, and I has since been extended to cover clothing 5 and footwear: and sundry other miscellaneous ' items." of general expenditure. Statistics" of. p-rices so obtained, and index numbers of the movements in food,

: rent, -and fuel and: light, have, been regularly', published, and the publication of index'numbers for all groups will be undertaken shortly.,.'., '" . .. : „' | •■' The data on which these' statistics are I based are obtained by the collection of | returns monthly and half-yearly from I retailers., in the various representative !■ towns covered ;by the", inquiry. Since there are differing prices for the same named commodity, the instructions to retailers, says the Statistician,, are that the "predominant" prices are to be entered in all cases—i.e.', "the cash prices of that

grade or quality-most frequently sold

I The returns, including ."rent," were col- | lected on this basis in the first two'years lof the collection. The relative import- ! ance of each commodity is maintained in ; -the aggregate by a\ system of weighing, | and in combining the index numbers for j tne respective towns their relative im- | portance is maintained by- weighing ac- | cording:-to population. .' .• . • i '.' A.DIFFICULT PROBLEM. i Now;'as far as the general run of conj, .fiumable articles, particularly foods such j as bacon; butter, cheese,-flour, oatmeal, I etc., is' concerned,' the ' 'predominant -| \ price" is easjly ascertained and returned, i but; in''thel case of: rent it is much more i difficult, if, indeed; not impossible,' to de-1 teimine the "predominant rent" for each ,j size of house.. The.modern tendency for ■j the luxurious : small. house or flat ■ "with

j'. every possible convenience and 'labour I saving: device and rnofor-garagerT-cdnse-p_ qnent,,no doubt, on• the limitation ;of r families, ; the" desrre for ease a»d com-'----3 fort,'and the difficulty"• of securing' efficiI enfc/ domestic '-: 'help-^has enormously Ini creased, the, range •. of ."weekly, rents'' ;.j being-:paid ;for/the/four, '.five,. arid:sii; ■j roomed houses or • flats; added to which :* there-is the^furtKer"cpmplicatipn'of situ- > atiori where/jtherinfluence of -land Values' j is.reflected iri'/fche rent£ paid.. :' -,-rV. '• .• To illustrate- the difßculties of statin" : predominant rents,: the Statistician' «up- •| P»es a table" sh6wing;-.the' .range' 'of-Tents 1 being.paid.for Jour, five,aHd>six-roomed ; ; houses, according to.the' Census'of 1921, ;■' }n the twenty-five representative <ow'as " covered by the rent'inquiry.. ■;' , *. „ "Ifc nas always, been the .aim of' tbe M'~ensus and, Statistics Office," says the' • Statistician,' "to avoid basing the' official statistics /on estimates, and to obtain j 'data representing actual cases wherever •j possible. ..Estimates, even by experts .] and when/most-carefully, madfe,. are sub'«j ject to human vagaries. "Very early the v conclusion come -to-was. that while 'preA dominant rents' -might:give approximate- ;-: Iv correct statistics-6f 'rents,' they .would f, not be satisfactory -for/measuring the ;.- 'movement' in *ent, which was the prini'i «pal object of ,the collection. In the >.early_rent returns, which were : closely ;;■ scrutinised,- it was. observed..,that • any V'movement shown was expressed nearly ;; always in fixed, amounts of Is in the case' £ of the lower.rents and 2s 6d'and--5s in j-,;-the case of -the higher Tents, and al/i though' provision, was made in the'return U books supplied to -retain copies of the [i returns,. so that the previous quotations >i should be available when making the new .•...returns,, .vet decreases were sometimes •j S hox™'- ■'■which -when queried were found ■j to he not intended..: After giving the „| matter consideration the : office decided >j that 'predominant', rents were not satisfactory, and if 'persisted N in wbuld-; in(i crease the- difficulty of ■ determining the 1] .truf>_rent-movernent and its extent.- It ; w,as further decided that the best pos--: jj:Bib.3 indication of the rent movement ■^ would be the difference between actual ;i rents paid from time to time. Follow- "■;. ing on this the present system of sample !j rents was introduced jn 1916. All re-. :.; Entered house agents, in .the twenty-five. jj towns covered, and as many owner rent- >/ collectors as could be ascertained, were ■ i required- to furnish each' half-year are- ■> i. turn giving for each size of h&use in -1 J-beir care-! the. actual weekly rents colj lectible and the: number of nouses cbvI'j ured. This'is the system .which is now Mm operation, and was'referred to as the ;<j 'sample system' in the December 'AbSj stract,' Ther.e can be no better methodffl of measuring /the rent-movement'than by •\ icomparing actual rents 'at one time with :/; actual,rents at another period, and there % * is no intention of'changing" the system." J QUESTIONOOrPr THE SAiIPLE. ' >■ . The number, of, four, . five, six, .and V seven-roomed -houses covered by the DeJj partment's inquiries fluctuated- slightly -] each half-year (going up to 11,436) until U August. 1919.. <it which ..collection it num-. j ■ij bered 10,777. From August, 1919, it fell C| rapidly '. until. February, 1921/ when .it | .-jl'siood at 6603, recovering, in August foi- * lowing to 6924, from which date it has :l

fluctuated between that and 6000, until August,: 1923. Tho size., of the sample has been sufficient in. normal times to give very closely. thY true- average rents —in. 1921 ,'it 'covered 19 per cent, of the. four, five,' and six-roomed housss in the twenty-five'iowns' dealt, with. The times, however, were not normal, and for the reasons explained in the December 'Ad-

stract' the sample lost to a certain, extent its 'representative, character by. nut including asutficient-number of the higher rented houses. , On' different occasions

£ sincje 1921 attempts1 have been made to ,\ obtain a larger sample; and, .-while"'not 'i much ..success has .been "~ attained, the % number has not been allowed to decline. -■\ A small . samp.le','. if representative; is sj just as''good'as a "large one, but is less '■[ likely lo be as "representative. A buc~r. cession of big decreases 'or big increases '- in the number covered is not desirable. ...A table 1 showing the : average weekly £'. house-rents as disclosed by the Census ;':: of April,--1921, compared with the\ half- * yearly |-collection of August, 1921, con- •^ tains the.followirig particulars relating to /■"the-four cEief centres-.— . ■ i

;-•] Auckland.-—Fotuvrooms': 'Census, 18s '.'j 4d; half-yearly collection, 14s 32d. ;V| Five rooms : Census, 21s sd; half-.-'.■j yearly coyection, 19s id. Sis"' rooins_: »;,. . _' ■ . ■,;.■ ■•:.;-;.'. '. >

Census, 245-3 d; half-yearly collection. 23s 4id. Seven' rooms ;! (Census, 25s 2d; half-yearly-collection, 29s 6id., Wellington.—Four rooms :■ Census, 20s 4d ; half-yearly collection, 15s B£d. Five rooms : Census, 24s Id ; half-year-ly collection,. 19s 4^d.' Six rooms : Census, 27s 9d; half-yearly collection, . 24s 9d., Seven rooms : Census, 32s 9d: half-yearly collection, 32s 2d. . Christchureh.—Four rooms : Census, • 35s 6d; half-yearly-collection,; 13s Sd. Five rooms : Census. 20s 4d; -halfyearly collection, 19s 2Jd. Six rooms : Census, 21s IQd; half-yearly collection, 26s sd. Seven rooms : Census, 25s lOd; half-yearly collection,- 27s lO^d. Dune.din.—Four rooms : Census, 13s 3d; half-yearly collection;' 12s 5Jd. Five rooms : Census, 16s 8d ,• half,yoarly collection, 15s l^d. Six rooms V Census, 19s 5d •-half-yearly collection, 20s l|d. Seven rooms : Census,. 24s 4d; half-yearly collection, 24s 4£d.

The point to be noted in connection with these averages is that they are the arithmetic mean, and that the influence ox tile high rents included is- to make the average consideraoly higher than is actually being paid for the majority of the rented dwellings. The influence of the higher rented houses'on the average is considerable, although their numbers may-be" few: For instance, the average weeKly rent of all ■ jfour-roomed houses at the Census of 1921 was 16s 6d; only 2.19 per cent, of the total were at '45s and'over, yet if -they are omitted1 the average falls to 15s 7d; and; if those at 40s and over (3.16 per cent, of the total) are omitted the average falls to 15s""2d. Similarly, in the case of fiveroomed houses the average is 20s, but with those at 45s and over (3.44 per. cent, of the total) omitted it falls to 18s 8d: with those at 40s and over (5.93 per cent, of the total), omitted it is 18s Id. ' v For four,, five, six, and sevenroomed houses combined the average is 19s -10d: with those at 40s 'arid over (6.16 per cent, of the total) omitted it becomes 17s : lid.-

■The weighted average: weekly rent for four, five, six> and seven-roomed houses as shown- by the half-yearly collection in August, 1921, was 19s lO^d, whereas ..the actual average weekly rent." as show^n by the Census in April ,'of that year was 20s 7fi. Notwithstanding the .number of cases in which the half-yearly collection figure was less than the Census, the average ;so ascertained was in 72 per cent, of the cases represented by' v figure as high /as the average actually paid .for the majority of the houses iv these cases. This can perhaps be more readily seen by taking the weighted average of the 'four, five, six, and sevenroomed houses in the'four chief urban i areas ;— . . . - . .

' Proportion . percent. : . of Houses -,- • as shown m . • by Census Weekly." Weighted April, ■ Average ". ,' Average 1921, at Kent TJrban Area. Weekly }ient, or below as • . , ■ " ■ . Half-yearly Kent ili -shown by ■ Collection, / previous Census of '' \ • Aug., 1921. Column. Apl., 1921. .' „'■' . . .-- s. d. Percent; s. d. " AncWand -....-.. 20 i 62.66 21 6Wellmeton- ..:. 21 6 52.51. .' 24 4 . Christchurch .. ' 20-10 "'•' 72 62 ' 19 0 ' Dunedin ...... 17' oi' '.67.91 . / 10- 3 V THE AVERAGE /RENT. ' ■-_ From ihe point of view of the rent index for the Dominion, the sample average rent of the four, five/ six, and seven-roomed - houses in the twenty-five representative towns, says Mr. Fraser, is ■properly comparable with- the average Ayeekly rents disclosed by the Census for all sizes bf houses-'iii all cities and boroughs, for which 'the sample' is taken as "representative. On this basis the com.paiison, quoting the half-yearly collection immediately'before' and "after the; census, is'as follows.:—Half-yearly collection; February, 1921 .(Dominion ..weighted average), 18s 6i'd; Census^-all sizes, all /cities, and.-.boroughs, 19s. 9d;, half/yearly collection^"August,'l92l,' -19s: lOid. > Bearing: in/mind;;'thkt.. thesejfig-' uresi.tepresent i the arithmetic mean, an attempt has.been made to ascertain from the Census figures the "medium"* and the .'."mode,".and these are found.to.be approximately 17? 6d and 20s respectively. The average, weekly rent of four, five,. _ six,. and seven-roomed houses in all '[cities and boroughs according to the Census was 19s 10d.' .:'■■■.; .. . . ; • .|

_It is pointed out that in 191 applications made to the Labour Department during 1923 under the renfcrestrictkms legislation, the average Weekly rent in 1914 was shown, as 14s IOJd, and in 1923 21s',9d, representing an increase during the period of 46 per cent.'- The rent index in February, 1923, showed an increase over .1914 of 43.7 per cent., and. in August, 1923, of 49.8 perceni,. In the 191 cases the rent demanded averaged 27s lid, being an increase'of 88 per cent, over 1914. In 127 cases the rents were adjusted by the Labour Department or a Magistrate, the average rents' in these • cases being :14s 3|d in 1914 and'2ls Hid adjusted, representing an increase of 53.42 per cent., or 3.62 per cent, -more than the rent index for August, 1923.' ;In 104 cases of four, five, six, and seven ■ rooms, the average rent in 1914 . was. given as 15s , 3|d,. .while : the' rent actually being paid- in' these cases in 1923 averages 21s 9|d, an increase of 42.45 per cent, as against the rent index increase of 43.7 per cent.' in February, 1923, and 49.8 per cent in August, 1923. ■; Of the cases adjusted 80 referred to houses of four, five, six, and seven rooms, which were '■ adjusted by the Labour Department or a Magistrate from an average -of- 14s-9d in 1914 to an average of 23s 2^d, representing an increase of 57.34 per cent. These" cases were apparently under Jthe average in 1914, which accounts for the large- increase. The - half-yearly collection gave the average 'in 1914 at 15s s£d and in August,. 1923,- it. was 23s l|d, which is only id below the average at which these were adjusted in. 1923. ... "In conclusion," says the Statistician, ' "the whole matter may" be summed up by saying that in 1921' the statistics of average rents ■as ascertained by the half-yearly collection for the respective towns, and particularly so far as four and .five-wwrned houses are cpn-' cerned, were understated, but the average so ascertained for the'whole Dominion 'was very close'to the actual average as disclosed by the census. While there is. no definite test of the rentmovement afforded by the census,'it may be fairly claimed that what-evidence is available from - census and elsewhere) tends; to-support the conviction that the rent index does accurately represent the rent-increase, which has on the average W re^'t d houses taken P*ace since 1914. . Althqugh while the. sample was falling in 1919-21 the upward movement, was slightly retarded, any understatement was only temporary, > and the loss was. subsequently mostly, if not wholly, recovered. -It. may be pointed out that it-is not possible to mathematically demonstrate the' accuracy or inaccuracy of:statistics oj this character' If two or- more sets of such statistics,', bearing ore the same- subject .but derived from different sources* closely support each other withqut actual > agreement, and there are only small differences, both may: be taken as accurate; one cannot be.said'to be absolutely correct and the other approximately correct—both are approximately. correct. As stated on a previous occasion, the Dominion rent index is a much better index than is generally conceded'by its critics, and certainly there is no better indication of the rent movement available." RESULTS OF FEBRUARY COLLEC- '■■• TION.._; . The collection just completed for February, 1924, covered. • 7W7 houses' and gives the Doniirubn -weighted averngb

weekly rent for each size of house included in the index as follows :— , Average „, .. , Number. Weekly Size of House of liouaes. Rent. 4 Rooms .'.... 2,972 ; 036 11 J • 2.479 1 3 1J 6 •» 1,768 1. 80J 7 -»■ /•■ 018 112 OJ ' Dominion weighted average 4, 5, 6, find 7 rooms com- - - b'ned : .. ( 7,537 - 1 4 1J The index-number' for February, 1924 based on July, 1914, is 1563, represent ing ■an increase of. 4.34 per cent over August 1923 (1498), and 56.3 per cent over July, 1914.