The Failure of Economic Forecasting
Introduction
The use of statistics in economics has had a bad press. “Lies damned lies and statistics” is an accusation often levelled at economists and politicians who use statistics to try to prove a dubious point in a debate. It is commonly asserted, “You can make statistics prove anything”.
One story demonstrating the mis-use of statistics is told of two economists who wanted to know why they were becoming drunk each night after dinner. They decided to use a statistical analysis in their enquiry. During dinner they consumed several Scotch whiskies with water and became very drunk. The next day they repeated the process. They had exactly the same food but switched from Scotch whisky with water to Irish whiskey with water. They still became drunk. The third night they repeated the process but this time varied the drink by having rye whisky with water. Again they became drunk. They concluded, that since water was the only constant factor consumed during their meals then it must have been the water, which was making them drunk.
Cynical anecdotes ridiculing the fatuous claims made by economists and statisticians reflect the poor track record they have in predicting trends in the economy. In THE DEATH OF ECONOMICS (1994), Professor Paul Ormerod criticised the statistical performance of organisations like the Economic Co-operation and Development organisation (OECD) and the International Monetary Fund (IMF). The statistical models used by both organisations failed to predict the Japanese economic depression of the 1990’s, the economic recovery of the US in the second half of 1992 and the turmoil in Germany after unification with its unemployment level of 4.4 million unemployed workers. It has also failed to anticipate the crisis in the e-commerce sector of the economy and in the telecommunications industry with their tens of thousands of job losses.
Professor Ormerod compared the forecasting methods of both the OECD and the IMF with one devised by himself by using a simple benchmark comparison. The benchmark he used for his survey was a “naïve projection that next year’s growth of output or inflation would simply be equal to this year’s” (p 105). The benchmark used required no knowledge of economics to produce and a forecast could be made by “anyone who understood the elementary arithmetic of percentage changes”.
Professor Ormerod went on to say:
Over the 1987-92 period this extremely simple rule performed at least as well as the professional forecasters in projecting next year’s economic growth rate. And in terms of inflation, the rule performed as well as the OECD and IMF, and slightly better than the national governments. In other word, the combined might of the macro-economic models and the intellectual power of their operators, whether based in national governments or installed in tax-free splendour at public expense in Paris or in Washington, could not perform any better than the simplest possible rule which could be used to make a forecast (p. 105).
To add to the discomfiture of economists Bill Jamieson, economics editor of the DAILY TELEGRAPH, pointed out;
“The bad, sad truth about economics is that it is ill suited to forecasting and the record of the treasury’s own economic forecasting bears this out. Not only has it consistently misread the pace of growth in the economy but also it has failed to predict any one of the major recessions since the end of the second war”. (AN ILLUSTRATED GUIDE TO THE BRITISH ECONOMY p.15).
Mr Jamieson concluded;
It is not that government economists have several buttons missing on their calculators, rather that their keypads are passed in a quite random way by agencies outside their control: external shock, or what Harold Macmillan was wont to call “events, dear boy, events”.
The “external shock” Marx would have described as an economic crisis, a feature of the anarchy associated with commodity production and exchange for profit.
Forecasting: Getting it Wrong
Statistics must be handled with care and treated with caution. Statistics are filtered through a political process often to be used by one group of capitalists and their politicians against another. Statistics by themselves do not make the case for Socialism. They should support an argument and not be used in place of one.
Marx’s CAPITAL is an excellent example of the use of statistics to illuminate and support his arguments against economists like Say, Malthus and Senior. Socialists are aware that capitalism produces statistics for the benefit of capitalists and their governments, and not for the working class. Statistics for Socialists are hard to come by. How many people, for example, actually live off the unearned income of rent, interest and profit and in what percentage of the population do they represent? We would like to know but no research grant is likely.
Misleading statistics have been published on productivity and unemployment which clearly over-estimate the productive power of capitalist production and the Cassandra-like pessimism of rates of unemployment associated with increased out-put and the impact of new technologies like IT and Computers.
During the past 200 years the use of power driven machinery and other developments in the techniques of production have increased the productivity of the average worker. All agree that this is so, but there has been continuous disagreement about the rate at which it has increased and is increasing.
The Government Statistical Service compiles and publishes an index showing the rate at which “output per person” increases from year to year. This is calculated from estimates of the total volume of national production each year, divided by the average number of persons employed during the year.
It shows that between 1966-73 productivity rose by 32% per annum, between 1973 and 1979 it rose by 1.2%, between 1979 and 1983 it rose by 2.1% and between 1983 and 1988 it rose by 1,6% (ECONOMIC TRENDS). The average productivity increase for these years is 2.12% whereas during the 19th century the annual increase was also generally between 1% and 2% a year. The fluctuations in productivity are related to the trade cycle with depressions acting as a downward pressure on productivity.
A number of people assert that the rate of productivity increase in the future will be many times greater than the past average of 2% a year, and it will result in the great majority of workers being unemployed in this coming century, this in spite of the fact that current unemployment in Britain has been falling.
If these prophets of gloom now appeared on the scene for the first time, what they said would demand serious consideration much like those who made fortunes out of the Millennium bug which was to affect computer systems on January 1st 2000. It is therefore important to remember that in the past 150 years there have been a dozen similar forecasts of what is going to happen in the future; every one of which has been disproved by events.
Examination shows that all these erroneous forecasts have been based on a fallacious method of measuring output and its increase along with the fluctuations in unemployment levels, and that those who now offer to tell us what will happen in the coming years, like the OECD and the IMF, have fallen into the same error.
The Fallacious Claims Made For Productivity.
In THE WEALTH OF NATIONS (Ch. 1, 1776), Adam Smith illustrated the increase of output brought about by the division of labour by what he saw in a pin factory. The work of turning wire into pins was divided into 18 separate operations, with different workers doing each operation. Adam Smith reached the conclusion-if it can be called a conclusion- that the division of labour had multiplied the output of each worker by 240, or even 480 times. He was unable to be precise about these figures because, as he admitted, he did not know whether one man doing all the 18 operations would produce one pin a day or 20.
Adam Smith, in his account of the pin factory, dealt only with the last stage of the operation, that of tuning wire into pins, and took no account of the fact that by far the greatest part of the labour required to make pins had already been applied in producing the wire, including the mining of the ore, producing the metal from it, producing the machinery, buildings and so on.
In his book THE GREAT WASTE (1933), John Hodgson quoted Adam Smith on the pin factory, though he altered it to read that the output of the pin workers had increased “by many hundredfold”, and Hodgson’s own forecast was that if avoidable waste were eliminated output would be multiplied by ten.
The story of the Pin Factory caught on and is still being reproduced in economic textbooks. It has had the unfortunate result of leading many people to exaggerate enormously the rate at which output per worker has actually been increased by the division of labour and other changes in the methods of production.
Karl Marx, in his POVERTY OF THEORY (1847), quoted from Proudron’s PHILOSOPHY OF MISERY, which had recently been published, the statement that between 1770 and 1840, the workers had multiplied their output 27 times. As the real increase of the worker’s output in that period had been that it had about doubled, Marx’s acceptance of Proudron’s statement calls for an explanation. It is that it was not until many years after 1847 that Marx completed the formulation of his labour theory of value, which provided a valid measurement for output per worker and its rate of increase or decrease through the trade cycle. Proudron and Marx had both made the same mistake as Adam Smith.
As an amusing aside, an anecdote is told of the late Lord Joseph, Tory Minister for Education under the Thatcher regime, being incensed on being told that students at a university were being taught Marx’s theory of value. It turned out that what they were being taught was Adam Smith’s, undeveloped, cruder and more primitive theory of value although the “Marxist” economic lecturer thought he was teaching Marx’s own theory and Lord Joseph’s informant could not tell the difference.
Three years after the death of Marx, Frederick Engels, in his preface to CAPITAL (1886), put forward the theory that Marx’s “cycle of stagnation, prosperity, overproduction and crisis” seemed to have run its course, and to have been replaced by “permanent and chronic depression”, with little or no further expansion of the market for commodities.
Echoing Malthus’s “law of population”, Engels wrote:
While the productive power increases in a geometric ration, the extension of markets proceeds at best in an arithmetic ratio.
He did not elaborate on his assertion and provided no evidence to support his belief that “productive power” was increasing at a rapid rate each year.
Engels’ formula meant that the rate of unemployment, which was 10% in 1886 (based largely on trade union returns of unemployment), would go on rising indefinitely. Events within commodity production and exchange quickly showed Engels that he was wrong. Engels had greatly exaggerated the rate of increase of productivity and was quite wrong about supposed failure of the market to expand. British capitalism in 1996 sold about four times the quantity of commodities it sold in 1886. He dropped his erroneous theory, as any scientist should in the face of contrary evidence, and reverted to Marx’s sound theory of the trade cycle with its period of boom, crisis, depression and up-turn.
In support of a campaign about unemployment the Independent Labour Party in 1909 published a pamphlet called the machine Monster. It contained the forecast:
Within the next decade…the productivity of the human unit will have multiplied a hundredfold.
And concluded:
A machine operative in an up-to-date, high-speed engine shop is equivalent (as a producer)…to fully 50 skilled workers of 30 years ago.
But this argument failed to allow for the additional labour required for the production and maintenance of the machines.
Capitalism took no notice of the I.L.P. Output per worker increased, as usual, between 1% and 2% a year and unemployment, which was 8% in 1909, steadily declined in the next 10 years.
When unemployment was rising in the 1930’s (the official figure reached 23% in 1933), Major Douglas and his followers were telling the Labour Party that it would have to change its name because there would soon be few “labourers”, only the unemployed. The prediction never came to pass.
In 1952 the Labour Party published TOWARDS WORLD PLENTY, in which it was stated that developments of machinery and technology from the industrial revolution onwards had raised productivity a thousand fold in the industrialised countries.
In fact the increase had been something like fourfold.
A year or two earlier, Ernest Bevin, Foreign Secretary in the Labour Government, had called on the workers to strengthen his hand by increasing production by 10%. The workers might reasonably have told Bevin that if they had already increased their output a thousand fold, why worry about a piffling 10% more.
In December 1954 the European Productivity Agency of the Organisation for European Economic Co-operation (OEEC), published a report on the TECHNICAL PROGRESS AND FULL EMPLOYMENT. It forecast “an unprecedented displacement of manpower”. And went on to say;
Some economists predict that within the foreseeable future the average worker will be able to produce at least the amount of goods now produced by five men.
Examples can be found where one worker operating the new equipment produces as much goods as 100 more produced before.
Taking 20 years as “the foreseeable future” (Tony Blair’s time period in which he claims he will end child poverty), and as the number of workers in employment was approximately the same in 1974 as in 1954, the volume of national production in 1974 according to the OEEC report, should have been 400 times the level of 1954. It actually increased by only 60% they were wrong too about their prediction of a huge “displacement of manpower”.
Unemployment in 1974 was still very low, at 2.6% compared with 1.5% in 1954. The OEEC failed to recognise that a necessary part of the labour required increasing productivity has to include the labour needed for the production and servicing the new equipment. How many businesses now employ an IT unit of computer specialists whose function it is to repair computers, which have failed to work and repair the network when it crashes? Then there are the services I.T, specialists who come in and up-grade or replace the existing hard or soft ware. And so on.
In December 1963, almost a decade later, a conference called to discuss the question of full employment met in San Francisco. Dr. Arthur Carstens, Director of Labour Programmes for the Institute of Industrial Relations, University of California, addressed the conference with these words;
I think we will in the next decade, learn to produce all the goods we need in the US with 2% of the working population.
He thought that the rest of the workers would have to find what he called “made work”, such as “selling second mortgages to each other, or engaging in psychiatric work”.
Other speakers at the Conference forecast a vast increase in unemployment. But ten years later unemployment, which had been 5.7% in 1963, was 6% in 1973.
A year later, in an article in THE PEOPLE (6.12.1964), Arthur Helliwell reported that in the US;
Experts estimate that robots are gobbling up jobs at the rate of 40,000 to 70,000 a week. This means that between 2 million and 3.5 million men are being thrown out of work every day in a country where there are already 5 million unemployed.
If those experts had been correct, unemployment in the US would now be between 68 and 119 million-it is actually 7,236,000 (INTERNATIONAL MARKETING DATA AND STATISTICS, 1998). What is forgotten is the fact that computers break down and need to be serviced. The computer sector has actually generated employment in its own right, from production through to exchange, particularly with the growth of home computers which increased by 25% from the 1980’s to 1996 (source: Salomon Brothers, July 1997). It is forgotten by many economists that commodities have to be sold and have given rise to large computer warehouses and outlet stores.
In his Reith lectures in 1965 the late Sir Leon Bagritt, Chairman of Elliot Automation, said;
In the United States…one man on the land produces more than enough food to feed fifteen men in the cities and in fact there is a surplus of food grown by this small proportion of the American labour force.
It is clear that he knew nothing about farming and food production. Most “farm” workers do not work on farms. Most of the labour required in the production of food is the labour of workers in industries supplying input for agriculture such as chemical fertilisers, weed killers, pesticides. Machinery, tractors, oil, electricity, cattle foods and so on, and in the canning and food processing industries.
The 1970’s saw these extravagant claims continue. There was a report by THE DAILY MAIL (16.7.73) of a statement made by Jack Peel, Director of Industrial Relations for the Common Market Commission, and formerly a trade union official. Mr Peel said; Well before the end of the century less than 50% of the population of working age will be working.
His prediction was utterly wrong. In the final decade of the 20th century there were in Britain some 27.8 million in work out of a working population of about 35 million.
Not to be outdone, Professor Stonier was reported in THE TIMES (13.11.78) as giving evidence to the Government Central Policy Review Staff in which he said;
Within 30 years Britain will need no more than 10% of its labour force to supply all its material needs.
In 1980, under the heading of “By 2001 only 1 in 10 may be working”, THE EVENING STANDARD reviewed a book by Professor Stonier. No critical analysis was given by the newspaper to Professor Stoner’s preposterous claims. If the forecast had been correct and if unemployment had been rising from the 6% of 1978 to 90%, unemployment would now be 8 million and rising fast. Fortunately for the Standard many of its readers would have never noticed having fallen asleep on a train journey back from work after opening the first page. It is in fact just under one million and has been falling for most months after the last depression.
Professor Stonier fell into the same error as Sir Leon Bagritt. THE TIMES (13.11.94) reported Professor Stonier as follows;
Professor Stonier says his analysis is based largely on historical experience. At the beginning of the 18th century 92% of the labour force worked on farms; today only 2% do so. In the United States about 3% of the labour force produce practically all the country’s domestic food needs plus substantial exports.
It is the elementary error of not comparing “like” with “like”.
In the 18th century a farm was almost entirely self-contained. Now its input is largely dependent on imputes from a number of other industries.
The Poverty of Forecasting.
So why did the forecasters get it wrong? There is a common factor in most of the forecasters disproved by events. It is that the forecasters looked at what was happening in the final stage of each industry’s production process and failed to take into account the process as a whole.
As, for example, looking at the labour displaced by a machine and ignoring the labour needed for the production, maintenance and operation of that machine.
It was left to Marx to provide a valid measurement of productivity and its increase by the application of his labour theory of value. In accordance with that theory the value of a commodity corresponds to the total amount of labour socially necessary to produce it. But, that amount of labour is not merely the labour needed in the last stage of production but the whole production, from start to finish. Marx wrote;
In calculating the exchangeable value of a commodity we must add to the quantity of labour last employed the quantity of labour previously worked up in the raw material of the commodity, and the labour bestowed on the implements, tools, machinery and buildings with which such labour is assisted (VALUE PRICE AND PROFIT, International Publishers 1976, p 32).
How important it is to make calculations on the “whole” labour and not the “last” labour alone can be illustrated from the bakery industry. Some years ago a writer, mistakenly, stated that if, through improved methods, the number of workers in a bakery was cut by half, that would halve the value of a loaf of bread.
By far the greater part of labour required to produce loaves of bread is the labour applied before the flour reaches the bakery producing the grain, milling it and transporting it, producing the machinery, the buildings, the fuel etc. It is probably not far wrong to assume that, of the labour of 100 workers needed to produce a given quantity of loaves of bread, 90 of the workers are “previous” labour, and only 10 are “last” labour in the bakery.
If the bakery labour is reduced from 10 to 5, the calculation of productivity based, wrongly, on the last labour alone makes it appear that output per worker has doubled, an increase of 100%. By the correct calculation, based on the fact that it now takes only 95 men to produce the quantity of loaves before by 100 men, average output per man has increased, not by 100 per cent, but by approximately 5%.
It is interesting to notice that the calculations of the increase of “output per person employed” compiled by the Government Statistical service, the Office for National statistics, are based on the same principles as that used in Marx’s formula.
However, it is feared that this will not prevent the continual appearance of false forecasts in the coming years of the new century, as in the past 150 years. In 1999, at the fag end of the Twenty First century, SBC Warburg, Dillon Reed, Liverpool Research and the Centre for Economics and Business Research all believed that unemployment in Great Britain would reach 1 million by the year 2000. They were wrong. It stood at about 1.3 million. But what is an odd 300,000 discrepancy over a large glass of whisky.
Conclusion.
Statistics and the statistics used to illustrate the failure of capitalism and its economists is not an easy subject. However a preliminary study of Marx’s text, notably LABOUR, VALUE AND PRICE will allow someone to grasp the subject matter and how it should be analysed and presented.
From a Socialist perspective the predictions of futurologists is largely academic. In 1900 unemployment was a major problem for the working class as it still was in 1904 when the Reconstituted Socialist Party of Great Britain (1991) was formed. Reconstituted Socialist Party of Great Britain (1991) came into being arguing that capitalism caused unemployment as it did other social problems facing the working class. The position of Reconstituted Socialist Party of Great Britain (1991) then, as it is now, is that, if workers do not want to be unemployed either through the introduction of new technology or a as a result of a periodic trade crisis and depression then they will have to first, consciously and politically, establish Socialism.
The real question, which workers should be considering as a matter of urgency, is not the silly predictions about productivity and unemployment but the creation of a social system, which has neither unemployment nor employment. An unemployment statistic hides real social problems and real hardship. Unemployment is not a natural problem but a social one deriving from the anarchy of commodity production for exchange and profit. Unemployment cannot be isolated from all the other social problems workers face. The cause has to be dealt with and that cause is capitalism.
Throughout the entire 20th century, social reformers have claimed that they would be able to resolve the problem of unemployment, and that common ownership and democratic control of the means of production and distribution by all of society was unnecessary. Unemployment is still with us. The social reformers have failed. Like war, unemployment as a social problem passes on into the 21st century. Do workers really want to endure another century of unemployment? That is the choice they have to seriously consider. Capitalism or Socialism?