COST OF LIVING.

Star (Christchurch), Issue 17246, 12 January 1924, Page 5

COST OF LIVING.

RENT STATISTICS.

Tho following supplementary pronouncement of the Court of Arbitration regarding cost of living (rent statistics) for the six-monthly period, April to September, 1923, has been filed. The Court considers it desirable, in view of the rent statistics of the 1921 census, which have recently become available, to make a further pronouncement of a supplementary and explanatory nature, .**>onie very important passages in the Government Statistician’s memorandum accompanying the census figures have apparently been overlooked, and an impression has been created that t.ho Court, when dealing with rents, has woiked on inaccurate data. Statistical science is a somewhat involved study, and presents many pitfalls to the inexperienced. We -therefore do not suggest that the conclusions of those who have in rather exuberant language, expressed the opinion that the census statistics have shown the Court’s rent, figures to be hopelessly wrong, are dishonest or intended to mislead the workers and the public generally, hut, it is obvious that their conclusions are based on unsound reasoning and an imperfect knowledge ot the facts, and require correction. To begin with, the statistical returns commonly known as “ tho Court’s rent figures” are the Government Statistician’s official rent-index statistics, based on returns, collected half-yearly, of actual rents. As we explained in our last pronouncement of November, 1923. the system adopted is the sample system ; that is, particulars of the actual rents received in lespeot of a number of houses let by land agents and private owners, but principally bv the former, are ascertained at six-monthly intervals, and the percentage increases or decreases are calculated from tho returns so collected. Wo said. “ The total number of houses covered is 6.039. It is freely admitted that, owing to the abnormal housing conditions prevailing, even so large a number of houses us 6039 ma\ not he fully representative, and that the average rents received for ihos© houses may be less than the average rents received for all rented houses.” The Government Statistician has now' been able to check the average rents as shown by the half-yearly rent-index returns with the results of the 1921 census; and as was anticipated a discrepancy has been disclosed. This docs not, however, mean that a similar discrepancy exists in the percenLage increase shown by the half-yearly returns, and it is on the percentage incicases and decreases that the Court has always based its wage adjustments. The Government Statistician points out, in his memorandum, that “ although the discrepancies in the two sets of figures indicate an understatement of the averagfc weekly rents in 1921. this does not necessarily mean that the index of the rent-movement compiled from the halfyearly rent returns is also under-stated. It is not possible to get a proper comparative check on the movement from the census figures, the particulars available at each census not being quit© comparable : but it is to be noted that the rent index number does show a very substantial increase in recent years; in August, 1916, the rent index number baaed on 1914 stood at 1008, and five years later, in August, 1921, it was 1286, an increase of 27.57 per cent during the census period; in August, 1923, it was 1498, or 49.8 per cent above July, 1914.” It is on this 49 8 per cent that the Court has based its calculations of the increase in th© cost of living in so far as it is due to increased housing a sts. As the Government Statistician has pointed out, the census statistics for October, 1916, and April, 1921, are not strictly comparable. The former are based on rent® and rental values, and the latter on rents only, and the areas covered are not identical, though ap proximately so. Still, a comparison is of some value as affording a rough means of checking the results of a comparison of the half-yearly rent index figures for a corresponding period. The increase between 1916 and 1921, as shown by the census returns in respect of the thirteen metropolitan and suburban areas, was 24.45 per cent. The increase between Debraary, 1917, and August, 1921, as shown by the halfyearly rent-index figures prepared from the half-yearly colleetiqns immediately following the respective censuses was 26.20 per cent, or 1.75 per cent greater than that disclosed by the census returns. There has been no census since 1921, but the half-yearly rent-index figures for August. 1923, show au increase over 1914 of 49.8 per cent, or a further 21.2 per cent during tho last two years. This percentage is in accordance with general observation, and is certainly not in conflict with the census returns.

The census returns. as already stated, show that the actual average of all rents is higher Ilian the average of the rents of the houses included in the half-yearly rent-index returns. This is due to the “sample” (6039 houses) having ceased to be properly representative. The obvious explanation is that the houses now placed in the hands of land agents for letting are mostly of the older and poorer type, and it is principally from information supplied by agents that the half-year-ly returns are compiled. This does not affect the minimum-wage labourer, for he would naturally select the cheaper type of house in 1914 and in 1923. It may. however, affect the shoo assistant and the tradesman to some extent, for a percentage of this class of workers occupy houses of a hotter type, and a larger “sample” including a greater number of houses of a better type, might show a somewhat larger percentage increase than that shown by the present half-yearlv returns. The Census and Statistics o&ce is endeavouring to obtain a larger “sample” for the February, 1924 collection, but in the meantime Die Court is of the opinion that the discrepancv. if any, must be small, and can be set against the Is per week that we decided, in November last, to keep in reserve in view of contingencies. In this connection, it is desirable to bear in mind the following comment made by the Government Statistician in his memorandum on the rent figures:—“ln recent years the, character and quality of the smaller houses have very greatly improved . Whereas in former years the four. five, or six-roomed houses were invariably occupied by the wage earners, to-day, owing to the pressure of the domestichelp problem, they ore frequently occupied by professional and business men. but the appurtenances, conveniences, quality, style and finish of the bouses are vastly improved.” The “sample” small house has deteriorated, and the average small house has improved in quality and its comparative rental value has been enhanced. It is therefore desirable that the sample should be made at least as fully representative of t-ne houses occupied by workers generally as it was in 1914. and the Census and Statistics office now has the matter in hand. It is necessary to refer to the reasons for the adoption of the .sample system. First, it is impracticable to take a complete rent census everv six months, and in anv event, if it were possible to do so, the returns would

not bo available in time to be of prae- * tical use. Secondly, the sample system is accurate and reliable, provided the sample is periodically revised and kept comprehensive and representative. Thirdly, it avoids almost entirely the j complications du© to sub-letting and | the erroneous statements made in the J census returns. For example, many j cases have been discovered of the rents i of comhined shops and dwellings being j returned as rents of dwellings; and of j the rent of a furnished house being j returned as the rent of an unfurnish- ! ed dwelling. Although the census | papers are carefuly checked, and every ! effort made to eliminate such errors, j it is impossible to he certain that ail | .Save been eliminated. Fourthly, estimates and opinions as to the average rents are unreliable and vary greatly. Fifthly, it would he impossible to expect to obtain reliable Dominion r© turns of as large a number as 6000 houses from private sources. Rents statistics have proved a source of difficulty and misunderstanding in other countries as well as in New Zealand. Shortly after we issued our last pronouncement, wo received a copy of the report of the South Australian Board of Trade on its last Cost of Living investigation, in which it was stated that the board was satisfied, after full consideration, that tho only evidence on which it could safely rely in computing the movements of rents was that of the Government Statistician’s figures. The board compared rent, in relation to general domestic expenditure, to a bulky and incompressible package that had to be fitted into a jportmartteau. Other items .ould be squeezed and adapted, but the bulky package had td be given its full quota of space. Th® comparison is not inapt, for rent is the largest single item of expenditure in the domestic budget and must always be provided for in full. Though rent has increased to a less extent than any other of the groups except food, that make up th© cost of living, the increase in rent is ; more noticeable because of the large ness of the single payment. All the j cost of living statistics are based on : averages, and rent shows a greater j range of extremes than any othei , item. Some rer.t-payer a pay little or no more than they did in 1914, while. 1 others pay greatly increased rents .but s no system of averages can avoid being satisfactory to those at one end of the scale and unsatisfactory to those j at the other end. We deem it necessary to refer to a statement frequently made in Court, and as frequently refuted, namely that the Court allows for a standard rent of 15s 3d per week. The Court does nothing of the kind. The statement in question is another instance of misunderstanding statistics. A collection of domestic budgets made in 1911 showed that the average amount expended on rent was 20.31 per cent of the total domestic expenditure. This percentage has been taken by the Court as a basis. It is, however, only an average. Th© labourer in 1911 probably spent more than 20.31 per cent of bis earnings on rent, perhaps 25 or 30 per cent, and th© tradesman probably spent less, than 20.31 per cent on rent. The so-called standard rent of 15s 3d per week has evidently been arrived at by taking 20 per cent of a labourer’s wage of £3 16s Id. This is an altogether unjustifiable misuse of statistics, on two grounds: First, 20 per cent of the minimum-grade labourer’s wage cannot by any stretch of imagination be regarded as a standard rent for all workers; and secondly, as already mentioned, the minimum grade labourer’s expenditure on rent was probably 25 or 30 per cent of his earnings in 1911 and in 1923. Yet, if we departed from the average in th© case of the labourer we \yould have to reduce rather than increase his wages, for a larger percentage .allotted to rent would have the effect of reducing the total expenditure necessary to maintain the 1914 standard of living, because rent has increased by less than 50 per cent, while the general cost of living stands at 57.5 per cent above that of 1914. Every increase shown in the rent-index returns has been applied, half-year by half-year, with strict mathematical accuracy, to the average percentage of the domestic expenditure allotted to housing-cost, and the method of computation has frequently been explained. We hope that this further explanation will settle the matter. In conclusion, we desire to refer again to a matter that we stressed in our last pronouncement. We pointed out that the term “ rent ” in the narrow meaning of the weekly or annual sum paid for the use of a dwelling was a misnomer. The Court deals rather with the average cost of housing, the term “rent” is accordingly to be given a wider meaning, covering not only rent in tho narrow sense, but interest on capital invested and on unpaid purchase money, and other outgoings, in cases where the owner of a house is himself the occupier. The 1921 census returns showed that only 53.96 per cent of salary and wage-earners were rent-payers, the remaining 46.04 per cent having purchased the houses they occupied. Interest, rates and insurance premiums have not increased to the same extent as rents, so that by adopting the rent-mdex figures as indicative of the increase in general housing costs we have given the average worker a considerable advantage. In summing up the position in November last we said, “Th© use of th© rents statistics in order to arrive at the movement in the average cost of housing is justified on foui grounds: (1) That if we disregarded the rents statistics w© should have to resort to the opinions or estimates of experts, which from the nature of the case would be most unsatisfactory ; (2) that, though not fully representative, the statistics available are accurate as far as they go; (3) that- there is everv reason to believe that they indicate with reasonable correctness the percentage movement in rents generally ; and (4) that, even if they understate the increase in rents in the narrow sense, this is at least balanced by their being used as an index of the movement of the average cost of housing in tho wider sense.” There is nothing in the census figures to cause us to alter the opinion expressed in the passage quoted. We now know approximately the extent to which the houses covered bv the halfyearly rent-index returns are not properly representative, a discrepancy which the Court discounted bv treating tbe movement in the rent-index I number as applicable to housing cost generally': and the census figures, so , far as they are available for com- | narison show ar. increase approximat ing to that shown by the index figures, ( thus indicating that the non-repre-j sentative character of the sample could have affected the accuracy of tbe rentindex only slightly, if at all. In order to remove anv suggestion that tbe conclusions of tbe Court ar© at variance with those of tbe Government Statistician we desire to place it on record that we have fiullv disooscod with that officer, in whom we have the i,ifcmo«t jV>© statistical matters doolt with in this pronoun'-’©moot, pod That he hp.s expressed th° view that the rent-index fiomros are th© T'est measure obtainable of the rent movej inenfc, ami indicate the movement with much greater accuracy than is generally recognised.