EVEN AS AMERICA is dismantling its highly centralized phone system and promoting competition, the Soviet Union is heading in exactly the opposite direction. It has lost patience with the welter of special-purpose and unconnected networks that it has accumulated and has decided to develop a single, centrally planned, centrally managed system with similarities to old Ma Bell.

The irony is that for once, the Soviet penchant for centralization may work for them. In fact, the phone system they are heading towards will likely perform the functions the Soviets want better than our system will perform the tasks we have in mind. What's more, not only will the Soviets benefit from their own centralization, they will benefit from the competitive world we are creating by having more sources to buy from.

During a recent visit to the Soviet Union, I had a chance to look carefully at their telephone system. Three important points emerged:

*The Soviets aren't always as backward or sloppy in technology as we may think in the West. Their competence is merely uneven. The Chernobyl disaster gives us one part of the picture. The efficiency of their planned telecommunications system portrays the opposite.

*The highly centralized Soviet system allows them to use advanced Western technologies efficiently, so long as they don't have to innovate. The Soviets can build a modern telecommunications system, so long as they can continue to buy Western technology.

*The driving force in Soviet telecommunications, as in many other areas, is security and miltary usage -- not economics. To the Soviets, the notion that a federal judge and a bunch of antitrust lawyers could demolish a militarily efficient system must be incomprehensible.

Soviet telecommunications is no small system. Telephone exchange capacity has increased by about 50 percent every five years since 1950, to a level of 20 million lines in 1980. At that rapid growth rate, capacity would reach 30 million lines this year. That's a lot of lines in a country of 220 million, the vast majority of which do not have access to local phones. In fact, this pace of growth is approximately that of the United States, although it is only a third of the quantity.

Granted, the Soviet system, telephonic as well as political, is vastly different from ours. Since it is a closed society, the Soviet Union does not even want to encourage free and easy communication between private individuals -- much less such fancy stuff as call-forwarding, 800 numbers, or nationwide rent-a-car reservation networks. The system will never carry a tenth of the private traffic the American system must. For that matter, the phone system the Soviets are headed toward would not work well if the economic system were liberalized even as much as China's. If factory managers were allowed to call 10 different suppliers to get competitive bids, the system couldn't handle it.

But for what the Soviets want -- a system geared to institutions, not individuals, which puts a premium on security and hierarchy, which can be used by both the military and civilians with the military always able to take priority -- the Soviets' proclivity for centralization is, for once, finally serving them well.

The noncompetitive Soviet economic system is inherently inefficient. But it turns out that if you want to support such a system with a communications network, telephony is one of the very few areas of modern technology in which it makes sense to have a highly centralized authority with huge resources and czarlike control. The Soviets may even wind up with less obsolete equipment left over in their system than ours contains.

The Soviet approach, of course, assumes that the Soviet Union will continue to be allowed to buy sophisticated Western communications technology on the open market. But that's a pretty safe bet. The ideas behind the devices may be high technology, but the actual objects -- from computerized switches to fiber optics -- have become so widely manufactured as to be virtually unembargoable.

The new technologies driving telecommunications involve sophisticated satellites, fiber optics, microchip-driven switching systems, and new methods that allow conversations to be transmitted as a digital computer code, rather than an analog of a voice.

In the laboratory, if not in the field, the Soviets are at the state of the art in most areas of satellite technology. They have developed an acceptable capability to produce microcircuits themselves. They are actively engaged in fiber optics research and development. And they have been very active and rather successful in developing their own programs to run on computerized communications components readily available from Western firms such as Sweden's Ericsson.

But technology is not the only factor driving Soviet telecommunications. Several years ago, the Soviets woke up to the new proliferation of special-use circuits being built independently by various ministries and institutions. The universities in Moscow were developing a local communications net to tie their computers together, for example, while the military was developing its own long-haul administrative telecommunications independent of anybody.

For both political and economic reasons -- a fear of giving any institution an independent communications power base and a desire to save money -- the Soviets told the ministry of communications to develop a single, all-encompassing, centrally planned and managed communications system.

At that point, the Soviets had to start making decisions about what kind of system they wanted.

Probably half the long-distance lines of this system will continue to use conventional copper cable. This technology has the disadvantage of not having very high transmission capacity, but does have the advantage of being very difficult to tap.

And security is supremely important for the Soviets. This is why they are so uncomfortable with the fact that they rely on land-based microwave relay systems much more than we do. While microwave is the cheapest way to link low-density long hauls, the signals can be easily intercepted. This is why most of the system's long-haul expansion will be based on satellite channels.

In the past, satellites offered major security problems to the Soviets. Both in the West and the East, the original philosophy was to keep communications satellites relatively cheap and uncomplicated, while keeping the complex, expensive part of the link in a very few ground stations that could be easily fixed if something went wrong. The problem with that approach was that simple satellites broadcast weak signals over broad areas, which laid them wide open to unauthorized interception.

The trend now is to build satellites that can transmit to the booming number of cheap ground stations that can be slapped on roofs. In order to do this, the satellites are being built to transmit powerful, highly focused beams. Such a signal is much more difficult for an enemy to intercept. It's almost icing on the cake that these new satellites provide for highly desirable all-digital, high-capacity transmissions.

The new system will provide a very high degree of encryption and security even, between two non-military ministries. In the United States, such calls would be in the clear.

The quality of Soviet civil telecommunications slowly will improve. Long distance telephony in the Soviet Union already is much better than most Western visitors realize. Local service also will improve, although my own guess is that the actual rate of modernization will be slower than planned.

I am much less confident that the Soviets will develop a satisfactory data communications capability, primarily for use in computers. That is an area in which innovation, rapid response to customer demand, and a very high level of experimentation and investment have proved important in the West. The Soviet system, with its rigidities, its centralization, its abhorrence of the messiness uncoordinated, competitive, small-team research and its tradition of ignoring the wishes of its users, is badly designed to adapt to modern data communications demands.

In short, the Soviet Union probably will achieve its plan for an integrated, mostly digital telecommunications network by the end of the century. Centralization, for once, will pay off in technological progress. The irony is that the same centralized system that will get the network built will make it difficult for the Soviets to take advantage of the network's ability to help them pull closer to the West in computers.