ESPIONAGE / SPYING BY THE NUMBERS : CIA Misread Soviet Economy : Agency’s dependence on computer models--instead of observation--produced results that were wildly askew.

For most of the past four decades, the CIA has been badly overestimating the performance--and prospects--of the Soviet economy.

The reasons may provide critical lessons for the future. In a nutshell, critics say, the CIA was led astray by the American bent for high technology: It relied too heavily on computer models and not enough on old-fashioned, first-hand observation and reporting.

In the early 1960s, for example, the CIA said the Soviet economy was already half as big as America’s and growing more than twice as fast. In reality, the Soviets were beginning a decline that has led them to the brink of collapse.

“Common sense was just not applied,” a senior defense official says.

By contrast, many of America’s allies trusted more to first-hand observation and their experts’ own judgments. On that basis, analysts in Sweden and Britain have long regarded the Soviet Union as essentially a Third World economy--an “Upper Volta with missiles.” But their views were dismissed in Washington.

The Background

The idea of using econometric models for intelligence estimates--which paralleled a general trend in economics at the time--was the brainchild of Harvard University’s premier Soviet economist, Abram Bergson, who touted the technique as more precise than the old system of sending in agents.

The trouble was that using computer models forced the CIA to rely on Soviet statistics, which were known to be flawed--or sometimes even fabricated.

Intelligence analysts tried to factor out the flaws in Moscow’s data, but the Soviet statistics were so inflated that they were almost useless.

CIA officials apparently could not fully comprehend the lengths to which Soviet officials went to pump up the numbers. According to one horror story that eventually emerged, finished ball bearings, instead of being used in the manufacture of new machinery, were sent back to Russian steel mills, melted down and re-cast, thereby boosting the mills’ production figures.

The Consequences

The consequences of the U.S. intelligence community’s preoccupation with scientific analysis may have been enormous. In hindsight, the early CIA figures gave undue credence to Nikita Khrushchev’s table-thumping boast in 1961 that the Soviet Union would “bury” the United States by the mid-1970s.

“We might not have turned ourselves into a debtor nation,” says Sen. Daniel P. Moynihan (D-N.Y.), or gone into Vietnam or embarked on the huge military buildup under then-President Ronald Reagan, if U.S. leaders had realized that the Soviets could not sustain their aggressive stance.

To be sure, not everyone is convinced of that.

Although the CIA chronically exaggerated the Soviets’ economic performance, it also underestimated the amount going into military spending, which some experts argue canceled out the errors on the overall economy.

Paul H. Nitze, who has served almost every President since Franklin D. Roosevelt, does not believe that the CIA figures on Soviet defense budgets had much effect. “After all,” Nitze says, “we were counting Soviet missiles and tanks with our satellites, and then responding to the threat-capabilities of those weapons, not to how much it cost to produce them.”

All that is argument about the past. What remains disturbingly true today is that U.S. intelligence methods used to measure the economic strength of a dangerous rival were badly flawed and that little has been done to improve the system.

In Senate testimony recently, CIA officials recognized that the agency has a problem, but they said they had made no major reforms yet.