Why did the Soviets not shoot down US spy satellites during the Cold War?

There is one other factor at play, which isn't mentioned in the answers so far. If you don't actually want a war it is advantageous to have mutual spy capabilities so you know in broad terms what the other side is doing instead of letting paranoia rule.

So in that sense spy satellites are an excellent tool because they provide information without having any offensive capabilities as opposed to, say, a spy on the ground which may also be used for sabotage.

Since both sides recognized this (don't have a concrete source but read it in several books about the cold war long ago), they both found it to their advantage not to shoot down the satellites even if they had the ability to do so.

Edit: Just saw that Andrews answer covers that aspect too.

I wanted to add something not covered in any of the other excellent answers:

A spy satellite does NOT just spy on one country, by necessity they all spy on The Entire World.

...

For all practical discussions, it is a physical requirements that a low orbit satellite intended to spy on a high latitude country like the USSR or the USA (less so, but still) will have to be in a polar (or near-polar) orbit whose period will be about 1.5 hours. Thus one satellite can and usually does cover all of the non-polar oceans and lands on the earth (in truth they likely cover the poles too).

So US/USSR spy satellites do not just spy on the USSR/US, they spy on everything, including themselves. Conversely this also means that to stop them spying on you, you would have to take out all of their spy satellites, not just specific ones.

Here's how that changed the calculations:

1. We could take out all their spy satellites. This would:

• remove their ability to spy on us,
• and also remove their ability to spy on everyone else.

This is all good so far, but note that the first, original intended effect was much more valuable to us than the second, unintended side-effect.

2. Obviously, they are going to retaliate and take out all of our spy satellites. This would:

• remove our ability to spy on them,
• and also, remove our ability to spy on everyone and everything else.

Note that in this case the second effect is as costly or even more costly than the original effect.

It in effect would reduce, by orders of magnitude, our ability to spy on countries and conflicts around the world that we, as one of the world's superpowers, need to be fully informed on so that we can properly attempt to influence and control them. What's happening in Cuba that no one's told us about or is trying to hide from us? What about Angola? Vietnam? What about places we haven't even thought could be a problem yet?

With spy satellites we have an instantaneous ability to focus on anything anywhere in the world with only 1.5 hours notice and NO DEPLOYMENT. Without them, we are blind by comparison and everything that our enemies and our allies do is a surprise to us.

For all practical purposes, to be a Global Superpower for the last 70 year, you NEED those spy satellites.

So in that light, this would have been a terrible trade-off for the USSR (and for the US), so there's no way they could have ever decided to make that trade-off.

Even if we simplify out the Kessler syndrome problem (it is not that bad in low enough orbits), we still have the economy problem.

A launch of a whatever "satellite killer" missile is of a comparable expenses to launching a new satellite.

Soviets could pretty much kill a single satellite, should they think it is important. Or maybe even few of them. If there was an ongoing war, this may be the right thing to do.

Outperforming the US space industry, on the other hand, was not really an option.

There is actually a legal-political answer to this question.

First, sovereign national airspace does not extend indefinitely. The Outer Space Treaty makes "outer space" belong to all mankind.

Article I

The exploration and use of outer space, including the moon and other celestial bodies, shall be carried out for the benefit and in the interests of all countries, irrespective of their degree of economic or scientific development, and shall be the province of all mankind. [...]

Article II

Outer space, including the moon and other celestial bodies, is not subject to national appropriation by claim of sovereignty, by means of use or occupation, or by any other means.

This treaty was ratified by the Soviet Union, United Kingdom, and United States in 1967. There are currently 112 parties to the treaty, including Russia and all of the other spacefaring nations.

Several equatorial nations did make an attempt to claim their airspace all the way out to geosynchronous orbit. The attempt failed, further supporting the concept that outer space belongs to all humanity instead of being the sovereign territory of a nation.

Second, it is true that international law has never formally defined exactly what "outer space" means. The Outer Space Treaty and other international treaties do make it clear that "the moon and other celestial bodies" are part of outer space, and they are even given specific protections. Nuclear weapons are prohibited from Earth orbit. However, these do not preclude other regions from also being including as part of outer space.

The preamble of the Outer Space Treaty begins

The States Parties to this Treaty,

Inspired by the great prospects opening up before mankind as a result of man's entry into outer space,

Because the treaty acknowledges that humans had already reached outer space, it creates a strong argument that anything at or beyond the regions of human orbit is also outer space. And spy satellites do indeed have higher orbits than manned missions.

Third, the Outer Space Treaty assigns jurisdiction of disputes to the United Nations Committee on the Peaceful Uses of Outer Space. Would a shootdown have occurred, it would be up to this body to decide if the incident actually occurred in outer space, and whether the incident was allowed or prohibited by the treaty.

This committee is political in nature, filled by member states with varying alliances and agendas. There was no guarantee that the Soviet Union would get their way in this committee. And even if the committee exonerated a Soviet action, the ruling could have given the United States a green light or a loophole to perform similar actions against Soviet satellites.

There simply was too much international political risk is performing such an action.

Let's rephrase the question:

Why did the Soviets not arbitrarily invade NATO territory?

The answer, of course, is "because that would have started World War III". And thus the question as asked, is answered.

What you are actually asking is, "why did the shootdown of Gary Powers' U-2 not start WWIII"?, and the answer to that is simple: the U-2 overflights were a very obvious violation of the USSR's internationally-recognised territorial sovereignty, and as such they were acting completely within their internationally-recognised right to protect that sovereignty when they shot down foreign objects over that territory.

In other words, the shootdowns of US aircraft over Soviet territory were purely defensive actions justifiable under international law, whereas shooting down a US satellite in what is effectively the neutral zone of space could only be construed as an attack.

Now, the USA could very well have attempted to use the U-2 shootdown as casus belli for attempting to start WWIII, but that attempt would have gone down like a lead balloon with the other members of NATO because the US, not the USSR, was very obviously the one in the wrong.

It's been alluded to in other answers, but a major problem with shooting down satellites is the satellite is going to fragment and these fragments will fly in every direction. Those pieces of space junk might seem small but they are moving at tremendous speeds (the International Space Station, for instance, moves around 17,000 miles per hour). In other words, let's say you blow up a US spy satellite. The pieces of that satellite might not just fall into the atmosphere and burn up. In fact, they might remain up there for years. Those pieces might hit your own satellites, which could in turn his more satellites, etc. This is known as Kessler Syndrome and was depicted in the movie Gravity.

We don't have to theorize about what these pieces can do either, as one hit the International Space Station robotic arm (image source compiled from NASA images)

In other words, you could shoot one down, but it might render orbital space unusable for you as well for years to come. It would be a very pyrrhic victory.

You do not shoot down satellites, you destroy them in-situ.

That results in a large debris field orbiting the planet that gradually spreads out.

Anything passing through that debris file will likely be destroyed and add to it.

You have no ability to ensure that your own satellites do not get damaged by your own actions.

There have been satellites that were destroyed and just about everyone involved in space were concerned.

The precedent for allowing overflight was first set by the USSR in 1957, with Sputnik I. From then on, neither major power saw it as to their advantage to challenge it - they were both keen to get their own surveillance up and running.

In 1957, of course, stopping a satellite was not possible, which may have factored into this decision. But by the time plausible ASAT technology was developed, space was widely used for surveillance and communications purposes by both sides, and the value of being able to use it yourself far outweighed that of trying to stop your opponent doing so. ASAT weapons were developed as a capacity to use in wartime, not as one to use to escalate a peacetime dispute. Both sides were aware they would lose their own capabilities quickly if so.

The logic of cold war deterrence also supported not trying to disrupt satellite surveillance. Suddenly moving to shut down the other side's satellites would be a direct attack on their strategic capacities and would be seen as a very escalatory move, plausibly the first step towards a surprise attack. This was not something you wanted to suggest to the other superpower, for fear they would move quickly towards "start shooting" in response.

Similarly, while actual aircraft overflights would be refused as showing too much, both sides benefitted indirectly from having some things be visible about their forces, using the "national technical means of verification" alluded to in SALT. Allowing the other side to have some limited sight of your forces, on a strategic level, is a stabilizing force - they know your fleet has not suddenly put to sea, or that you have not built another fifty missile pads, etc., and so they do not need to overreact to what might otherwise seem like small threatening moves. You're safer as a result.

And, finally, there are other options beyond shooting down satellites. If there is something specific you really don't want them to see, you can predict that the satellite will come over at 9.32am precisely, and an army can rustle up a lot of men with tarpaulins on short notice...