How did they foresee the Soviet collapse?

I’ve been rooting around in pre-Cambrian analyses of the Soviet Union’s economy. As all good Marxists know, once your economic base has cracked, the super-structure will not hold. The excellent work of Chris Miller here tells us that while Party Secretary Gorbachev knew full well his country was in trouble, his Dengist reforms failed, utterly. The US applied some pressure, but ultimately the Soviet economy collapsed, like any good blackhole, in on top of its own mess.

China’s economy is so different - freer, larger, more open to the world, pretty vibrant along the coast, less addicted to military spending, abundantly consumerist, way tech-ier. Kids line up for bubble tea, not bread, smartphones in hand. And, critically, it’s more know-able too - there’s plenty of (albeit questionable) data, you can walk around the shops (or rather sit at home and surf Tmall), rifle through listed company financials, hang-out at dinner with NDRC-ers and steel bosses. None of that was very possible in the Soviet Union.

And it seems to me there are more brains now focused on understanding the country. When I arrived in China all those years ago, there were only a handful of analysts who could tell you what’s what. (These days, when I’m feeling particularly miserly, I wonder if the analyst community could do with a sharp dose of “supply-side reform”. But then my better angels sing to me, and I know the truth (and the best analysts) will out.)

That said, we also know there is plenty we do not know - those known unknowns.

Here’s not the place to analyze what went wrong with the Soviet economy; that’s been covered excellently elsewhere (e.g. in Anders Aslund’s incredible output over the years). Today, I’ more interested in figuring out how the analysts who successfully figured out the collapse future of the USSR did so.

There are, I think, at least five lessons.

1. Focus on the important question(s).

In the Soviet case, it was: how big is the economy really; how much is being spent on the military and the empire overseas; and how vulnerable did that make the system?

Amongst other things, Andrew Marshall at the Department of Defense’s legendary Office of Net Assessments (ONA) obsessed about getting the size of the economy right, and figuring out how big the defense budget burden on it was. He was frustrated with the CIA’s estimates of the size of the economy (he suspected them too big), and estimates of military spend (he thought them too small). (The CIA did not have an easy job; even Gorbachev was not allowed to see the military budget in his first few years.) Marshall, affectionately known as Yoda, had a stab at both, and concluded, brilliantly,that the burden was extremely heavy. This insight was the lynchpin for those who wanted to defeat Moscow via a spending war. You can read about the late Marshall in Krepinevich & Watts’ The Last Warrior.

In Rubles, Dollars, and Power Austen Long at RAND tackles the debate about whether the CIA foresaw the collapse of the Soviet economy. He concludes that the agency did, eventually, see the stagnation, although they took a lot of prodding (not least from Marshall). Ronald Reagan noted in his diary in 1985 that Moscow’s economy was a ‘basket-case’, partly on their intelligence. But the CIA did not foresee a collapse.

Which leads us to…

2. It is extremely hard to call a qualitative change /structural break with any confidence. But you have to think outside of the box and lay out coherent future scenarios.

Force yourself to think about these possible futures by writing them out.

This is not an exercise in pure imagination - the scenarios should hew as close to what is possible, based on what you know of how the incentives are stacked and the system functions.

Published in 1989, Gur Ofer in Soviet Economic Growth: 1928-1985 discusses this issue, having accurately identified the economic slowdown which took place in the 1970-80s, as the capital-intensive, heavy-industrial growth phase ended. He notes that most studies estimated a 3% baseline growth path in the 1990s, via “muddling through”. But he also writes “we have to bear in mind that that its rare for analysts to foresee radical change”.

Indeed, it is. So we need to use our imaginations. (And admit that China is navigating that shift away from heavy industry with some success - ergo, all those lines outside the bubble tea shops.)

3. But that’s not an excuse for laziness.

Beware armchair economists, and those dependent on reading the English-language media. Focus on those doing the real hard work, the analysts who are walking the streets, interviewing folk in Chinese, diving into (or even better, creating) new data sources. Marshall partly relied on Soviet emigres who understood how the system worked from the inside. Some China hawks have called ten of the last zero economic collapses in China. One might sympathize with some of their political views, but one has to conclude that their understanding of China’s economy is deficient. Marshall always pushed his analysts to do the hard analytical work.

They’ll always be a market for chancers and grifters who’ll punt a radical change story; the challenge will be in sifting through the stories and figuring out what story makes most sense with the empirical data that’s available. Just saying “all the data is made up” is not good enough.

4. There are no facts about the future. And there are things about the present which we are never going to know with much certainty.

So put numbers on your scenarios (because your “quite likely” might not be someone else’s) and list the indicators you are going to track. Then periodically think about how your thinking about the probabilities have changed.

And this advice comes straight from the CIA! In 1983, Director William Casey ordered this Assessment in which he asked how well the Agency did predicting Vietnam, China’s nuclear test, the Shah’s fall in Iran etc.

Surprisingly, the agency gave itself passing marks on Vietnam. However, and this is an important point, they admitted they’d over-estimated the efficacy of US actions against the Vietcong and under-estimated that of their opponents - a bias which no doubt echoes in Washington today.

On other predictions, they owned up to missing the boat. And they too concluded their analysts had a bias to just extend the status quo.

So, think hard about what it is you are going to track to see how the scenarios you have laid out are going. And then periodically adjust those numbers as the evidence changes/emerges.

Personally, as in stock investing, I find it really hard to resist changing my probabilities when there’s a short-term wave of evidence contrary to my initial assumptions. I’m not sure I’ve figured out how to resist, but going back to those indicators sometimes help.

The British, we are told in the CIA assessment, have a word for the opposite of this error - ‘perseveration’: allowing an early judgement to blind you to change. Again, hopefully well-chosen indicators might help.

5. Discuss your views with other experts as much as you can. Do that like Amazon.

This one should speak for itself - they’ll challenge you, remind you of things you’ve forgotten, tell you you’re being dumb. And just the exercise of presenting your views helps figuring out whether they make sense.

If only conferences did much more of this and much less of the shoot-me-know, endless thick-pile-carpet presentations. Smaller, more focused groups work better - especially if everyone can take a page out of Amazon’s how-to-run-an-effective-meeting book and force everyone to read a short document before focusing on questions and debate.

There you go. Much of this is common sense. But I’ve found it useful at least.