International Historical Statistics

The principal sources of population data are official censuses, administrative enumerations, and registration records. For many countries in Africa, and some in Asia, these only began in the twentieth century, often only after the Second World War; and the quality of their coverage has not always been particularly good. Historical coverage of different countries, therefore, varies widely in scope and reliability. Table Al gives the statistics from all the censuses and official enumerations which have been conducted in these two continents and Oceania which were in their intention complete.¹ But since to restrict the table to these alone would provide no data at all for certain countries, and very little for others, the results of certain more-or-less official estimations have been included where these seem likely to be reasonably reliable.

The statistics in this section cover a wide range of topics and come from a variety of sources. The occupation data up to 1966 were largely derived by Professor Bairoch and his colleagues from national censuses of population, and are therefore rather limited in coverage for Africa and parts of Asia. The problems of accuracy referred to in the last section appear here also. But a still more significant difficulty is the very considerable variations which have occurred in classification, both between countries and over time. Professor Bairoch’s group refer to ‘the frequent changes in criteria and methods used in census taking’, and say that ‘it is practically impossible to come up with statistics that are perfectly comparable in time and space’.¹ The best we can hope for, therefore, is that the figures in table B1 are usable as a guide to structural changes within countries and for rough international comparisons.

The collection of some kinds of agricultural statistics began at a comparatively early date in some of the countries of Africa, Asia and Oceania. Indeed, statistics of grain production for Japan can be derived from the records of a survey conducted as early as 1645. These, along with the results of similar surveys in 1697 and 1829 are reported in the footnotes to table C2, but they need to be regarded with considerable caution. Sugar statistics for Mauritius and Reunion are available from early in the nineteenth century; South Africa, Indonesia, Sri Lanka, Australia and New Zealand — as well as Japan — were recording certain data by the middle of that century or soon after; and numerous other colonial territories had joined them before 1900.

Apart from table D1, which gives indices of aggregate industrial production, and table D13, showing the number of cotton spindles, all the tables in this section consist of physical output and external trade volume statistics for major commodities which possess an adequate degree of homogeneity to allow aggregation and meaningful comparisons between countries. Inevitably these are mostly basic commodities rather than finished goods — yarn and cloth rather than clothing, metals rather than machinery. The picture they give of industrial development is necessarily biased and partial. However, until very recent times there is little that could be added to them in the way of continuous statistical series for industries making finished products.

Because it has long been an important source of revenue to governments, external trade provides more statistical material at an earlier date for most countries than any other economic activity. It was always one of the first things to be recorded by colonial administrations, and as independent regimes were increasingly drawn into world commerce they began to follow West European practice in this field.

Governments have generally been more intimately involved in the provision of means of transport and communication than in agriculture or industry, at any rate until very recently. As a consequence there is usually more statistical material available from the past than on most other economic activities. Shipping was always a matter of close concern to major maritime powers, though there have been few of those in the continents covered in this volume. The railways were frequently of military as well as economic importance, and they necessitated the investment of large lumps of capital, on which the social return seems generally to have exceeded by a considerable margin that which could be captured by investors. Outside parts of Europe and North America it was fairly unusual for private promoters to attract this capital without some assistance from governments. Some form of reporting to governments, therefore, was required of most railways. Postal services were long recognised as a government function. Telegraphs fell naturally into the same niche in most countries where they were not simply an adjunct of the railways, and telephones generally followed. And in most countries the potential influence of radio, and later television, led to their direct control by governments, at least initially and in part, and hence to some form of licensing system, which produced a statistical by-product. Taxation of motor vehicles has had a similar effect.

Financial statistics exhibit some very great contrasts as regards availability. Some were collected and published from a very early date. Others may have been collected, but have been regarded as state secrets. And yet others were not collected for a long time, either because they were regarded as private secrets, publication of which it was beyond the competence of the state to compel, or because there were no permanent financial institutions from which it was possible to request them.

There is a considerable array of material on prices available for most countries, even former colonies, for a comparatively long way back in time. Most of this, however, is intractable in the extreme, and it has been decided to include none of it in this work. Instead, there are two sets of indices, based on this material; but reducing it to some sort of readily comprehensible order. It was felt that annual average prices in local currency of (say) a kilogram of rice were unlikely to be of much use except to the specialist, and that a general indication of overall price levels would be more widely useful. Of course, there is oversimplification and distortion in the process, and like all index numbers, those presented here require careful interpretation. In order to achieve exact comparability of an index over long periods of time the commodities and their weighting must remain unchanged. However, in a changing economy first the weightings and then possibly the commodities themselves will cease to be appropriate representations of the quantities and things actually used. All indices, therefore, must compromise between continuity and relevance, and the more rapid is economic change the more frequent must be the breaks in continuity. In the period since the Second World War, the official indices in some countries have been changed very frequently indeed. In other countries and periods where commodity consumption patterns have changed very little, the indices have not changed either. In this way the danger of lack of representativeness has generally been avoided.

Of all the subjects on which statistical material exists, probably none shows less uniformity, both over time and between countries, than education. There is no universal definition of what constitutes a primary school, or a general secondary school, and even that of a university has shown considerable flexibility in the recent past. Moreover, there have been several major reorganisations of school systems in practically every country, and minor changes have been very frequent. Furthermore, the statistics of pupils have not always been collected in a consistent manner, even within the same schools system, and data for teachers has been so sparse and inconsistent for Africa and Asia that it has not seemed worth including. In addition, the date in the school-year to which the statistics relate has been altered on various occasions in most countries; the exact meaning of the data has changed, sometimes referring to all pupils on school registers, sometimes to those in regular attendance, sometimes to those present on a particular day, and sometimes to those present when an inspector visited. Sometimes, indeed, they may just be teachers’ estimates. Moreover, private schools have not always been included in the data collection process, and these have been of considerable significance in some countries, and it is not always clear whether or not adult pupils are counted in the numbers at school. And whilst the problem of changing definitions of a university have been largely evaded in this edition by covering all forms of higher education in table 12, ambiguities remain about certain sorts of vocational training, and even about whether teacher-training is counted as secondary or higher education. Nevertheless, for all the impossibility of making precise comparisons over long periods of time, and, even more, between most countries, when used with care these statistics do provide useful comparative material, even if only of a very rough nature.

Whilst it has been the general principle in most sections of this work, only completely ignored in section H, to prefer raw data to those which have been processed, there are no such things as raw national accounts statistics, and everything in this section is synthetic, the result of elaborate calculations by sophisticated statisticians. It is really impossible to summarise briefly all the complex operations involved, and the user who requires descriptions of them is advised to consult the latest UN, Yearbook of National Accounts Statistics, or, for more detail, their National Accounting Practices in Sixty Countries, and the IMF, Balance of Payments Yearbook.