Israel's top economic officials were utterly bewildered by the 20% jump in unemployment after a switch to the OECD harmonized definition. Since last weekend, when the change was revealed, none of the country's leading economists has been able to explain how a different measuring method jacked up the numbers. On Monday the head of the National Economic Council, Prof. Eugene Kandel, took the unusual step of convening the body to discuss the ramifications of the startling results.
The compliments showered on Prime Minister Benjamin Netanyahu and Finance Minister Yuval Steinitz over Israel's low jobless rate, compared to other developed countries, might be unjustified. "Could Israel, which boasts of the high quality of its workforce, have been mistaken for so many years about unemployment rates?" astonished cabinet ministers and other senior officials asked. The answer, apparently, is yes. Unemployment is a full percentage point above that reported just two months ago. The people who claimed for years that the Central Bureau of Statistics figures were too low were apparently right.
By the old method of measuring unemployment, the rate in the last quarter of 2011 was 5.4% of the civilian workforce, representing 174,800 jobless. Unemployment for all of 2011 was 5.6%, according to the bureau. But last Thursday, when it published its figures for the first two months of 2012 and pointed out that the method had been changed, the rate for January and February was 6.6% and 6.5%, respectively.
Such a mistake has profound implications for many economic parameters, including growth, employment, wages and structure. The newly-discovered higher jobless count might explain why there hasn't been any genuine wage pressure for several years, since there were more job-seekers than reported.
Top Bank of Israel officials who met with senior staffers from the statistics agency this week, in an attempt to obtain clarifications, say they have not received adequate answers. Both the central bank and the Finance Ministry are waiting for the bureau to complete an internal probe before deciding on any further steps.
Economist attach great importance to the meeting called Monday by Kendal with senior officials in the statistics bureau. Both the CBS and the NEC are part of the Prime Minister's Office.
The controversy has provoked a slew of pointed questions from experts, such as whether joblessness was the only area in which the statistics bureau has furnished erroneous information over the years. In fact, for some time now the reliability of the bureau's information on inflation and housing costs has also been challenged. Although these are measured using other methods, it has been suggested that perhaps a broader investigation into the workings of the bureau is necessary.
"The Central Bureau of Statistics started a new series using a different methodology and didn't explain how it converted from the old series," said Dr. Michael Sarel, head of the economics department at Harel Insurance Investments & Financial Services and a former director of the Finance Ministry's Economics and Research Department. "I tried building the new series using the coefficients appearing in their announcement but didn't succeed. I tried calling the number they provided several times and sending an email but never got an answer. There is no real reason for unemployment growing from 5.4% to 6.5%, or how increasing the number of surveyed localities contributes to the rise in joblessness. This is a crucial figure." Sarel admits, however, that 6.5% unemployment is still considered low in international comparisons.
CBS tried explaining to the public last week the large gap between its unemployment figures at the end of 2011 and the beginning of 2012, but to no avail. The agency said the method of measurement was changed and adapted to the OECD's measuring technologies, and the bureau's workforce survey switched from a quarterly to monthly basis while increasing the sample size.
The main changes in the measurement process were the inclusion of career and conscripted military personnel in the workforce and the expansion of sampling from 370 to 470 localities. These changes, however, can't explain the jump in jobless figures. Unemployment in the military is 0% so including this group in the workforce sampling could only lower the overall unemployment rate. And adding 100 new small and marginal communities couldn't cause such a major shift in numbers, regardless of how rampant unemployment was in those places. And even if Israel had clusters of joblessness, with over 15% unemployed for instance, it is reasonable to assume that the statistics bureau would have already known about them.
The old employment survey had its weaknesses too, and the periphery, which hadn't been adequately represented, is now properly reflected, economics and statistics experts in Jerusalem said earlier this week. But going and canceling the old survey results is still quite a stretch, they suggested.
'No errors in measurement'
The experts, who preferred not being identified, added that decades of old unemployment figures will need to be revised to match the new methodology. We need to see if the new (revised ) data sets reflect the business cycles over the years, they said, adding that if no correlation is found the bureau will need to go back and reexamine itself. The bureau says it checked updated data from January 2003 to September 2011 but we didn't see the data set to be assured it matched the direction of the business cycle, they said. We need to know, for example, where we were in terms of the business cycle and unemployment at the height of the last economic crisis.
One thing that's clear is that unemployment is on the decline according to the new methodology. Several months ago it was 7%, and it is now 6.5%. The experts said they expected the bureau to be able to provide full answers within a week or two.
TheMarker has learned that CBS was asked to simultaneously release unemployment data for a full year using both methods, but refused. It did perform a comparative test for the fourth quarter of 2011 but did not release the results to the general public, nor is it clear they were made available to economic experts outside the bureau.
Mark Feldman, the bureau's labor sector director, adamantly rejects any claims that unemployment measurements were previously wrong. "There weren't any measurement errors," he insists. "Things changed. There is a fracture in the unemployment data series. We introduced chaining coefficients to determine the jobless figures according to the new methodology."
Feldman stresses that there was no sudden switch between methods, as suggested. "We've been working on this move for six years," he says. "We conducted a seminar last April that was open to the public, where we mentioned the transition. We conducted further seminars in December and January. Anyone claiming that we didn't agree to publish the results of both methodologies simultaneously for a year makes the claim in hindsight after seeing the figures (last ) Thursday. We presented the new methodology to the Public Advisory Council for Statistics in special symposiums and nobody said the methodology wasn't correct.
"For decades Israel, like all developed countries, measured unemployment according to the methodology of the International Labor Organization," Feldman continues. "OECD countries also work according to the ILO methodology, which includes two ways of measuring unemployment: Joblessness in the civilian workforce and in the overall workforce. Some countries use one method and some use the other. In Israel it was decided in the 1960s to work according to the civilian workforce methodology. On joining the OECD we were asked to switch to the methodology of the general workforce, including the measurement of the military workforce."
Feldman also obviously feels there's a problem with the bureau's new stats. "Many changes were made in the method, and we are now checking all the figures, which will take a month or two to complete. The transition to the new methodology involved people from academic, government, political and public circles. We are conducting a thorough probe into the reasons for the discrepancy between the two methodologies. I can say right now that there is more than one reason but can't say what they are at the moment. We will shortly issue explanations on what happened and they will be made available to the public."
Want to enjoy 'Zen' reading - with no ads and just the article? Subscribe todaySubscribe now