Hedonic Indexing: Now We Can All Have Productivity Miracles!
October 31, 2000
prudentbear.com
It surprising that a dry, highly technical branch of economics has become a source of such vitriolic debate amongst central bankers, but such has been the case since the Federal Reserve began employing hedonic indexation when measuring the performance of the US economy. Whilst we were in the midst of booming stock markets globally, endless prosperity and the apparent repeal of the business cycle, few bothered to question the general methodology, or the controversial economics behind the application of the hedonic deflator in particular. Those who did, such as Northwestern University professor Robert Gordon (and a consultant to the Federal Reserve as well), were ignored by and large.
Over the last few months, however, the US economy has continued to attract much of the world’s surplus capital to sustain its economic boom, despite growing financial fragility characterised by an increasingly large current account imbalance and a record private sector deficit equivalent to approximately 6 per cent of GDP. A large proportion of this capital is said to be drawn to the country on the basis of America’s widely perceived productivity miracle. We are uncertain as to the exact reasons as to why the United States continues to attract such large amounts of capital inflows, but it is increasingly clear that Europe’s capital markets in particular have proven an excellent funding source for this private sector savings deficit. As Dr. Kurt Richenbacher has noted:
“The biggest single item among recorded US capital inflows, is, in fact, borrowing by corporations and federally-sponsored agency bonds [from the Euroland bloc]. And their borrowing effectively implies the conversion of euro into dollars. Together, they amounted to $255 billion in 1999 and to $71.7 billion in the first quarter of 2000. Quoting from a Commerce Department report: ‘US agencies, in particular, have sharply stepped up their sales in international markets as part of their continued effort to expand the frequency and size of their issues in all markets.’ Keep in mind, the latter number is not annualised. In other words, dollar strength is clearly borrowed strength, not structural strength.”
Richenbacher’s analysis is important for two reasons. One, it casts doubt on the generally held perception that most of the capital inflows now sustaining the US imbalances are derived from European businessmen and portfolio managers buying into the American new economy. More importantly, it highlights the unstable quality of these capital flows; the latest flow of funds data appears to illustrate (as Richenbacher points out) that America’s private sector deficit is being funded largely through debt, rather than equity, a much more unstable form of external financing. Perhaps as a frustrated response to this apparently misguided and destabilising migration of capital to the United States, a few European central banks are beginning to question the foundations of this so-called American growth and productivity miracle. Recently the Bundesbank has added its voice to this debate, indicating that the use of hedonic indexing has done much to add to the illusion of US economic strength (consequently, enticing US dollar flows and engendering corresponding weakness in the euro) and the relative weakness of German economic performance in comparison. The discrepancy stems in large part from the fact that the German central bank statisticians employ no hedonic adjustments to calculate improvements in productivity, according to the Bundesbank. By adopting such techniques, German investment would appear bigger and faster growing, argues the German central bank.
Now Britain has disclosed similar findings. According to the Office of National Statistics, although the computer industry makes up only 2 per cent of total output, if the US calculations were used in Britain, it would have a significant impact, raising total recorded industrial output by an additional 6 per cent over four years. And so begets another new economy miracle. By sheer coincidence, an election is thought to be a mere 6-12 months away now in the United Kingdom, so any “improvement” in the country’s economic performance will no doubt enhance the credentials of the New Labour government, as well as providing respectable economic cover for a grab bag of tax concessions in the upcoming budget statement by the Chancellor of the Exchequer.
Now that this methodology is in effect being legitimised by other central banks and official sector agencies, the genie is well and truly out of the bottle, however questionable the economics that lies behind hedonic indexation. Much has been written on the latter. We are merely focussing now on the political aspects of the change in statistics: it is indeed surprising that it has taken so long for other countries to perceive the political benefits one could derive from pumped up productivity numbers, particularly since there is no economic equivalent of an Olympic drugs testing agency, which could sanction countries that resort to “economics by steroids”.
Predictably, the new charge has now been taken up by Goldman Sachs. Gavyn Davies and his international economics team in London has used some of the new methodology now employed by the Federal Reserve to determine whether global economic growth has in fact been understated. The report notes that in the first half of the 1990s, Britain’s adjusted business sector growth was a mere 2.62 per cent a year, against 2.61 per cent under the old methodology. But in the second half, Britain did better, averaging 3.43 per cent (2.9% unadjusted), not far behind America’s 4.46 per cent. Only Japan gets a bigger boost from hedonic adjustments, but its average growth is only lifted to 1.9 per cent from 0.82 per cent. Growth in continental Europe remains sluggish.
This analysis may appear surprising until one considers a recent statement by Professor Gordon, which helps to clarify the apparent disparity between Britain and the rest of Europe. Gordon has made the point that “wherever productivity is surging today, it is narrowly related to computers”. His work shows that there is no evidence of “benefits of computers and other electronic equipment spilling over to the sectors that have heavily invested in them.” In other words, the productivity miracle stems from the production of computers themselves. Unfortunately, it appears that this increase in measurable productivity is itself a function of the use of the hedonic deflator. In his study, however, Gordon appears to accept the validity of employing this deflator as a given, but its use does create a huge historic discontinuity in the data pre-1972, where there existed no such deflator to measure other equally important innovations which enhanced human welfare, such as electricity or antibiotics. Only one other sector – automobiles – has been given similar treatment, and the quality enhancement calculations involved have nothing like the impact that the hedonic deflator in computers has. For other high tech industries, such as pharmaceuticals, biotech, or health care, advances which generate huge improvements in “hedonism” or welfare, are not reflected in American (or European) price deflators; there is no resort to the hedonic deflator which means that a wide spectrum antibiotic such as penicillin has been ascribed no more real value than the aspirin tablet which was administered for want of anything else. To introduce hedonics now simply puts a barrier between the pre-1972 and post-1972 data, which is why its use was resisted for so long at the Federal Reserve. The temptation to inflate one's numbers and output, however, has ultimately proved too great, which is why such inhibitions have now largely disappeared: first, the Federal Reserve broke down the barrier, and now the Germans and the British are considering the introduction of hedonic indexing into their own numbers so that they too can better adorn their economic statistics.
Of course, as Davies’ own work demonstrates, this “productivity enhancer” will not necessarily solve the problem of “low growth” for many countries of continental Europe, which in fact largely explains the relatively sluggish numbers produced in the report for continental Europe in relation to America and Britain. Computers are not produced in countries such as France and Italy, which is why the introduction of hedonic indexing has done nothing to inflate their numbers (although it may help The Netherlands and Germany somewhat). The hedonic deflator measures output of PRODUCTION, not consumption, so these continental European statisticians have effectively walked into a trap if they hoped to make their numbers look more impressive by adopting this methodology when measuring the performance of their respective economies. All they have managed to do is further legitimise a rather questionable form of economic modelling by the Federal Reserve.
Indeed, the limitations of the methodology are implicitly demonstrated by the conclusions reached by the Goldman report itself. According to the report, in the second half of the 1990s, Britain increased information and communications technology (ICT) investment by an average of almost 28 per cent a year in real terms. Britain, the report concludes, has been closer to the American experience in recent years than has been generally recognised. Goldman then goes on to argue that if the United Kingdom had a bigger information technology sector, the country’s growth rate would have been even closer to America’s, reinforcing the importance of IT production as a means of enhancing economic growth statistics when applying the hedonic deflator. As an aside, it also suggests that Japan’s economic state is even more dire than had been implied by the old methodology, given that the report’s application of hedonic indexation in the case of Japan appears to be the main reason why the study shows a marginally improved economic performance on the part of the land of the rising sun over the last 5 years.
By the same token, if one were to introduce hedonic indexing to emerging Asia, then it is clear that countries like Taiwan, Korea, Thailand and Malaysia (which produce loads of computers and computer peripherals) would likely show huge increases in growth rates, likely well in excess of 10-15%, given the pervasive role of IT production in their respective economies. This would certainly undermine Western claims about Asia’s “flawed economic model”, which would no doubt please the Asian authorities. But is it a true reflection of the state of economic affairs in these countries right now? However much the use of the hedonic deflator might help to enhance Asia’s growth statistics and thereby legitimise its economic model, these inflated growth statistics, if employed, would have little credibility amongst investors. The potentially higher GDP rates derived from the methodology do jibe at all with current grim economic realities in those countries chiefly manifested through the dismal performance of their respective stock and debt markets recently. It is noteworthy that over the past few months in particular, emerging market debt credit spreads have started to blow out again to their highest levels in over a year, local currencies are again declining sharply against the dollar, and the stock markets in countries such as Taiwan and Korea have fallen by over a third since July, the very sort of economies which should be “booming” if one accepted the legitimacy of the Goldman Sachs’ approach.
Let us illustrate the fallacy of hedonic indexation in another manner. If one were to imagine an economy for which the sole output was computers, then its GDP would be determined by the use (or abuse) of hedonic measurement. The well-being of the country concerned, however, would be almost entirely unaffected, because any apparent increase in GDP resulting from hedonic measurement would be offset by deterioration in the terms of trade. Thus, we would have an exact opposite of the situation in major oil producing countries such as Saudi Arabia, in which a rise in the oil price has no impact on measured GDP, (assuming unchanged oil production), but a massive impact on the Saudi economy.
Would it not, therefore, be easier to settle this debate about productivity once and for all by going back to a statistical methodology which doesn't introduce these multifold problems? It is a very dangerous road to travel down, given the extent to which such statistics might ultimately undermine the credibility of the government’s own statistics. If the Fed and its supporters are so truly confident of the assertion that the US productivity miracle is for real, let's see them demonstrate it under a less controversial methodology employed reasonably successfully by all OECD economies over the last several decades. Then we would indeed resolve this vexing debate once and for all. |
