We are all well aware of how the micro-processor and its applications in information technology have changed the way we work or play. Robots have changed fundamentally the process of extracting or converting raw materials and distributing the goods and services we consume in ways that astonish and amaze us. We worry about how they appear to replace people like ourselves in the work place.
The power of mobile devices to connect us to our customers, colleagues, friends and information and entertainment of all kinds grows continuously, as does our dependence on them. Young people live happily (we hope) almost exclusively in their cyber worlds.
But all this information technology is not showing up in productivity measures – that is output per hour of work – as one surely thinks it should or would. All those factories, warehouses, cargo liners or railroads and ports with fewer workers and ever more sophisticated machines and devices that support those with jobs, must surely raise the ratio of what is produced to the number of person-hours employers provide compensation in wages and benefits for. The numbers indicate otherwise.
Alan S Blinder, a distinguished academic economist, Professor of Economics and Public Affairs at Princeton University and recent former vice-chairman of the Federal Reserve, writes in the on-line Wall Street Journal of 24 November of The Unsettling Mystery of Productivity, with the sub-title: Since 2010 US productivity has grown at a miserable rate. And no one, not even the Fed seems to understand why.
Blinder refers to the available history of productivity. Over 143 years of records show that the US has increased measured output per person hour employed outside of the farms by an average 2.3% p.a. That is, output per worker has increased nearly 26 times since 1870. Clearly the more valuable output workers are expected to produce, the greater real benefits (wages) they may be able to earn from employers competing for their more valuable services. Between 1948 and 1973, described as the golden age of productivity growth productivity grew by an average 2.8% p.a – yet between 1973 and 1995 it grew by only 1.4% p.a on average. It then picked up again growing by 2.6% p.a between 1996 and 2010 only to slow down to a miserable 0.7% p.a on average since; for reasons nobody, according to Blinder, seems to know why.
In South Africa productivity as calculated by the Reserve Bank has grown, on average, by a mere 1.02%p.a since 1970 and by 1.92% p.a on average between 2010- 2013. But in the seventies the price of gold doubled and doubled again allowing the mines to profitably reduce the average grade of ore they mined and extend the lives of the mines More rock was extracted expensively from the bowels of the earth but less gold was produced with more workers – meaning lower productivity and much improved profitabilty. Since 1995 productivity in SA has grown by 2.8% p.a on average despite the recent slowdown.
The unpredictability of productivity matters to the Fed and the Reserve Bank because their task is to align aggregate demand for goods services and labour to their potential supplies, using the tools of interest rate settings and money creation at their command. Not too much and not too little demand is called for. Too much means inflation – too little means deflation, which is regarded as equally or even more dangerous to well being. But knowing just how much means being able to predict potential supply upon which productivity growth would have had an all important bearing. Such productivity forecasting powers seems unavailable, so greatly complicating the task of monetary policy.
The problem to my mind is a measurement problem. The issues involved in converting business revenues, measured in dollars of the day, into equivalent volumes that can be compared over time. Productivity is the ratio of real output, real volumes of goods and services produced and charged for, to the number of person hours needed to produce them. But how is one to compare the value of a good or service produced 20 years ago with its equivalent today? An aspirin produced then is the same quality as an aspirin taken 30 years ago. But the same could not be said of a life saving drug available today that was not on the pharmacy shelves 30 years ago. The quality of medical care, given these technological gains has increased almost immeasurably. What then does the so called inflation of medical costs included in the CPI mean when what is being paid and charged for are much improved medical benefits? You are not comparing like with like, apples with apples, aspirins with aspirins.
Nor can an off the shelf or off the internet personal computer or laptop today be compared with those of 30 years ago when access to the internet was first initiated. They have the computing power that would have filled a large office with mainframes 20 years ago. And the same could be said of television monitors or motor vehicles or so many devices that are incomparable in quality with the options available then, perhaps infinitely better given that the ordinary of today would have been unimaginable not so long ago. A similar observation can be made of a modern automated machine tool when compared to the machines utilised before.
Therefore, if we are to compare real output over time we have to allow for changes in quality in order to generate an appropriate series of prices and what indeed the benefits received cost the consumer. Prices have to be quality adjusted if any sense is to be made of the volume of output produced and measured over time. Volume of output calculated for the purposes of measuring real output for GDP or productivity estimates is revenue in money of the day earned by businesses divided by what is hoped is a realistic measure of prices. If quality has improved dramatically or indeed infinitely in the case of goods or services previously unknown, this price denominator, known by economists as a deflator (deflating nominal values into real equivalents) has surely to take on a very large number with a proportionately large impact on real volumes. The Fed is conscious of the danger of underestimating quality gains regards the inflation it targets of less than 2% per annum as effectively deflation.
Can we have any confidence at all in the numbers attached to deflators that reduce the revenues of businesses to equivalent volumes or convert nominal GDP with its real equivalent? I would suggest that we can very easily underestimate quality gains and hence over estimate the numbers called deflators. Quality adjusted prices may be vastly lower than they are estimated to be. If so volumes produced would be much higher and productivity gains much greater than estimated. The mystery to be solved is an appropriate deflator especially for goods or services with infinitely higher computing power and value to their users. There may in fact be much more deflation about than is recognized. Hence monetary policy may be even tighter than it appears.
The closest relevant deflator I could find was for the prices charged by US retailers of appliances and electronic goods. This deflator, designed to measure the volume of these goods sold by the retailers with a base of 100 in 2009, had declined to 68.9 or some 37% over four years. In the US, the prices of all retail goods rose by 8% since 2009.
The closest equivalent deflator provided by Stats SA was for Furniture, Appliances and Electronic goods Retailers that showed a decline of 8.5% since 2009 while all retail prices rose by 23% over the same period.
Are these deflators and all the others that convert value to volume accurate enough to form the basis for productivity comparisons? One must doubt this. There is clearly enough room for error to add an average one or two per cent per annum to measured productivity growth.
But while such uncertainty about the relationship between price and quality changes may bother the economists and the Fed, they will be of little interest to the firms that produce goods and services. They will be hoping to add to profitability by managing, as best they can, the relationship between revenues and costs measured in money of the day, including the link between the money of the day costs of employing labour and what each employee may be adding to the top line. In fact employing more, relatively unproductive labour, may well be the more profitable option, depending on their cost of hire even if such employment maximizing decisions reduce productivity. The South African economy would do better if firms were hiring more low skilled less productive workers rather than making the efforts they do to raise the productivity of much better paid, but relatively few skilled workers with advanced equipment and superior data management.
It is be noted that while productivity is seemingly in decline, in the US profits as a share of output are at close to record levels. The impact of innovation on productivity and GDP may be mysterious given the difficulty of devising a suitable deflator. The influence on profitability would appear to be unambiguously helpful for shareholders.
And consumers of goods and services (known and unknown in abundant quantity) can be comforted that excess profits tend to be competed away and they will pay no more than it costs to supply them, costs that will include a required return on the capital employed by competing firms. The objective of business and their owners is to maximise profitability, not productivity. Real output and so real productivity are artifacts of economists and statisticians, not businesses, for which profits and return on capital are the key measures.