#13 - JRL 9132 - JRL Home
Moscow Times
April 27, 2005
Real GDP Figures Are Anybody's Guess
By Alex Fak
Staff Writer

Number crunchers like to joke that economic statistics are like bikinis: What they reveal is important, but what they conceal is vital.

If so, then Russia economists have lately resembled prurient teenagers, straining their imaginations to figure out what exactly the State Statistics Service might be concealing.

The agency declared in February that the country's real gross domestic product grew by 7.1 percent last year. The figure, it turned out, was exactly the annual rate needed to meet President Vladimir Putin's goal to double GDP by 2012.

The problem is that official figures for real GDP growth are virtually meaningless, economists say. An unclear methodology, outdated weights given to certain parts of the economy and haphazard accounting for the so-called shadow economy all mean that the GDP figure here is more prone to political tinkering than in other countries.

In addition, the transitional nature of Russia's economy means that even honest attempts to estimate its growth are bound to fail, the State Statistics Service concedes.

As last year wore on, several signs pointed to a slowing economy. Monthly GDP indicators from Central Bank subsidiary Moscow Narodny Bank, which are used to forecast official figures, slid in the second and third quarters before dropping sharply in the fourth. The statistics agency's own figures showed a slowdown in growth in the third quarter.

State officials began talking down full-year growth. Just five days before releasing the unexpectedly strong 7.1 percent figure, State Statistics Service chief Vladimir Sokolin put 2004 economic growth at 6.5 percent to 6.7 percent -- below the government's official expectations.

Then, in announcing 2004 growth on Feb. 2, the statistics agency said it would start using a more modern methodology to measure growth in industrial production. Instead of adding product groups, like foodstuffs and machinery, the agency said it would switch to basing output on economic activity such as mining and manufacturing.

Since the new methodology simply involves breaking down industrial production into different categories, the overall number should not have been affected. But it was -- and with a vengeance. Industrial growth jumped from a preliminary estimate of 6.1 percent to 7.3 percent, and more than doubled in November alone. An apparent industrial slowdown in the second half of the year dissipated, and manufacturing growth actually accelerated under the new numbers. Overall GDP for the third quarter was revised from 6.4 to 7.1 percent.

Analysts, from academic statisticians to the World Bank, found the new data "difficult to interpret," in the words of John Litwack, World Bank's chief economist in Moscow.

In a semi-annual World Bank report released in March, Litwack noted in particular that "the much higher growth in machine-building in the new figures ... is rather puzzling, given declines in growth reported in previous quarters" and higher costs of metals in 2004.

Economists noticed other peculiarities, too. Inventories apparently skyrocketed by 40.5 percent last year -- "an unusual feat given that output of goods was reported to have decelerated significantly, while consumption accelerated," Troika Dialog wrote in a research note to clients.

Another concern involves the State Statistics Service's way of "deflating" nominal growth in the value of goods and services to weed out an apparent increase in production that is purely the result of higher prices.

Vladimir Tikhomirov, senior economist at UralSib bank, said the statistics agency assumed particularly low inflation in sectors like agriculture and trade, where information is collected by polling and is therefore hard to verify. That has allowed those sectors to post higher real growth. "Playing with numbers proved to be a painless option, particularly because it could be masked by a change in methodology," Tikhomirov said.

When methodologies change, government officials tend to publish both the old and new series side by side for a while. But the State Statistics Service did not do that, leaving economists guessing, said Al Breach, an economist at Brunswick UBS.

Fradkov in the Henhouse

Tikhomirov said the change of methodology appeared to be aimed at shielding Prime Minister Mikhail Fradkov.

"We believe that the changes were prompted largely by governmental concerns that falling economic output could have a major political implication for Fradkov's Cabinet," Tikhomirov said.

After a top-to-bottom government overhaul early last year, the State Statistics Service became part of the Economic Development and Trade Ministry, which is headed by German Gref, one of the government's leading economic liberals.

Critics, however, immediately discerned a potential conflict of interest, because Gref's ministry carries out macroeconomic projections and because the statistical data directly reflect on its work.

So Fradkov placed the statisticians under his direct supervision last April.

Since then, Fradkov has repeatedly upbraided ministers for their "lack of ambition" in forecasting economic growth.

During a particularly heated Cabinet meeting in August, he threatened to fire Gref, saying, "Not a single scenario in the [Economic Development and Trade Ministry's] forecast is optimistic or even aimed at the goal of doubling GDP."

"Why can't we put 7.5 percent annual growth into the forecast?" Fradkov said at that meeting. "It's not just anyone, it's the president who called for the doubling of GDP."

Statisticians, however, deny facing any interference from the state.

"No part of the government -- be it the Central Bank, a ministry or the presidential administration -- interferes with our work," State Statistics Service head Sokolin told reporters in January.

Andrei Kosarev, who worked for the statistics agency in the 1990s, said it would prove difficult to pad the numbers, even under state pressure. "GDP is an indicator that is intrinsically linked to hundreds of other openly published indicators," said Kosarev, now chief macroeconomist at the Bureau of Economic Analysis. "The huge volume of publicly available statistics would make any such attempt futile."

Yet shortly after Fradkov began supervising the State Statistics Service, statisticians at NTC Research, which compiles GDP data for Moscow Narodny Bank, began noticing a sharp divergence between their indicators and official figures.

For instance, against a slowdown in the industrial sector, global growth and higher oil prices, official statistics indicated that the market services sector was prancing at close to 10 percent growth -- twice what the purchasing managers' index was predicting, said Christopher Williamson, head of economics at NTC.

Another economist recalled that strong economic figures have helped the government save face before.

"This is just like two years ago, when the December inflation brought [the CPI index] just within the [government-set] target," said the economist, who is employed by the state and asked not to be identified for fear of reprisal. "Again, that was very convenient."

Some economists said they have been unable to replicate the State Statistics Service's calculations.

"We're trying to arrange a meeting with the State Statistics Service to talk about their methodology and to understand it, because it has been very hard to get any information from them," Williamson said.

He and other economists complained that the agency keeps mum about how it compiles key indicators, such as the exact contents of the consumer price basket that it uses to measure inflation.

Reached on the telephone last week, Irina Masakova, director of national accounting at the State Statistics Service, said she was tired of battling cynicism surrounding the 2004 GDP figures.

"All the criticism is unfounded. People who write it don't seem to know much about how national accounts are calculated," she said.

Quarterly GDP estimates are always up for revision because they are based on a preliminary sample of businesses, Masakova said. Rapid growth in the value of inventories could mean any of a number of things -- perhaps businesses moved most of their production but could not sell the newer, more expensive products -- and, anyway, inventories account for just over 3 percent of total GDP, she said.

Masakova noted that the statistics agency's complete methodology could be purchased from the agency's publisher. And, she said, counting GDP is mammoth work that often involves painstaking deductions of what is not observable.

"It is always easy to criticize," Masakova said. "But I could swear by anything you want that we are not fiddling with the numbers."

Higher and Higher

Gross national product, which is similar to today's GDP, was introduced in the West during World War II as a way to keep track of wartime production. It has since become the broadest indicator of a country's economic health -- for lack of any better measure.

GDP statistics in any country are plagued by problems. In the short run, the measure treats all monetary transactions the same, failing to distinguish between, say, an investment in a child's education and costs associated with crime, pollution or natural disasters. (In one well-known example, statisticians registered devastation and cleanup costs linked to Hurricane Andrew in 1992 as a $15 billion boon to Florida's economy).

Also, GDP only measures open market transactions, leaving out much of the subsistence production that is more prevalent in less developed countries such as Russia. For instance, some 90 percent of all potatoes consumed in Russia are grown by families for their own tables, said Kosarev, the economist at the Bureau of Economic Analysis.

So the State Statistics Service tries to estimate the amount of both the "hidden" economy -- characterized by tax avoidance -- and the "non-formal" economy, like barter trade.

The statistics agency puts the size of the combined shadow economy at 22 percent to 24 percent of GDP, Masakova said.

The ratio has held steady over the years, she said.

But the flat tax on corporate profits introduced in 2002 has pulled more and more businesses out of the shadows.

That suggests the statistics agency is double-counting many companies as both "open" and "hidden" and exaggerates GDP growth, said Anton Struchenevsky, an economist at Troika Dialog.

On the other hand, if the shadow economy, unencumbered by taxes, has been growing faster than the rest of the economy, some of the double counting might be balanced out, and GDP might not be as exaggerated after all.

Whatever the truth is, no one seems to know. Nor can anyone say for sure by how much the statistics agency exaggerates annual GDP growth by failing to re-weight the components of the economy every year. Essentially, that means the economy in 2004 was compared to a snapshot of how the economy was in 2000, despite vast structural shifts over the four years. Among other things, the shortcoming tends to exaggerate real growth in the state sector.

Using 2003 as a base year for comparison would bring economic growth in 2004 down to 6.9 percent, Struchenevsky said.

That is not to say that the statistics agency's mistakes are all biased in one direction. A common complaint is that the agency actually underestimates the size of the shadow economy. The estimated inflation for the value of inventories in 2004 was too high, opening the possibility that they have grown even faster than figures suggest.

Moreover, the government systematically undercounts small businesses, which tend to grow faster than large corporations, according to the Center for Macroeconomic Analysis and Short-Term Forecasting. Its analysts insist that industrial production grew in 2004 by almost a percentage point higher than official calculations.

Measuring the country's economy is tricky by its very nature, said Masakova of the State Statistics Service.

"We live in a transition economy whose structure changes rapidly," she said. "Only in more or less developed economies do you get to work with set numbers."