Why do booming economies stall?

The raise that never came

Picture opening a payslip after twenty years in the same line of work. The number on it is a little larger than it was two decades ago, and so are the prices, and when the two are set against each other the pay buys about what it bought back then. The raise, in any sense that reaches the shops, never came. This is not the story of one unlucky worker. It is the ordinary experience of a whole generation in a country that was once the fastest-growing economy anyone had watched, a place that rebuilt itself from rubble into the second-largest economy on earth inside a single lifetime and then, around the time this worker started out, quietly stopped climbing. The news still reports that the economy grew last year, by a few tenths of a percent. Standing inside the flat decades, a person could be forgiven for asking whether the growth is real at all, and, underneath that, a colder question: is the country finished?

That worker is Japanese, and the twenty years could be almost any recent stretch, but the question reaches well past Japan. Every economy that grows quickly will, at some point, stop growing quickly. This chapter is about what happens then: why the engine that produces a generation of fast growth can wind down, what the slowdown does and does not mean for the people living through it, and whether a country that stalls the way Japan did is, as the fear runs, doomed.

What a plateau looks like

Put Japan’s output on a chart and the shape of the thing is hard to miss. For three decades after 1960 the line climbs steeply. Output per person rose about fivefold, at something like five percent a year and faster still in the 1960s, as the country did what fast-growing economies do: it took technologies and ways of working already proven elsewhere and put them to use at home. Then, around 1990, the climb ends. An asset bubble that had lifted Tokyo land and share prices to famous heights peaked, shares late in 1989 and land about 1991, and burst. The machinery of how such a bubble inflates and gives way belongs to a later volume (Volume V, on how asset bubbles inflate and burst). What followed was not a crash in the level of output so much as a change in its slope. Across the three decades since 1991, output per person has risen by only about a quarter in all, well under one percent a year. The line that used to climb now barely tilts.

Here a distinction has to be drawn with care, because the chart and the payslip are measuring two different things. The line just traced is output per person, roughly everything the economy produces divided by everyone in it, and that number did not go flat. It kept rising, slowly. The thing that went flat, the thing the worker in the opener actually takes home, is the real wage, and Japanese real wages have barely moved since the late 1990s. A country can produce a little more per head each year and still hand its typical worker a paycheck that stands still, so long as the extra goes somewhere other than into wages. So a salary flat for twenty years is true of the wage and not quite true of output, and holding the two apart is the first step to seeing what the stall was and what it was not.

The engine that runs out

Why should fast growth run down at all? The answer starts with what makes it fast. A country far behind the technological frontier can grow by copying. It can import the machines, the methods, and the ways of organizing work that richer countries spent decades and fortunes discovering, and skip the discovering. Borrowing a proven method is cheaper and surer than inventing one, so a country with a wide gap between what it does and what is already known to be possible can close that gap in a hurry. The economic historian Moses Abramovitz described this catch-up as turning on a country’s capacity to absorb what it borrows, and built into it a catch: the advantage is spent as it is used. Every imported technique narrows the very gap that made importing so profitable. A country that has drawn level with the frontier has run out of things to copy, and to keep growing it must now do the harder thing the frontier countries do, which is invent what does not yet exist. Invention is slow, costly, and uncertain in a way that adoption is not. The engine that carries a country up was never built to run at the top.

Why the handoff failed

A frontier economy that had merely run out of things to copy would slow to the ordinary pace of a rich country, a percent or two a year. Japan slowed further than that, into a long stretch of near-zero growth and gently falling prices, and economists have argued for thirty years about why. Two explanations sit at the front of that argument, and they do not agree with each other.

The first looks at balance sheets. When the bubble burst, Japanese companies were left holding assets worth a fraction of what they had paid, while the debts they had taken on to buy them stayed exactly as large. Richard Koo argued that firms in this position stop behaving the way the textbooks assume. Instead of borrowing to invest, they bend themselves to paying down debt, and they keep at it even after the central bank has cut interest rates to zero, because the trouble is not the price of borrowing but the hole in the balance sheet. When a whole country’s firms pay down debt at once and almost none of them borrows, the money that would have been spent or lent drains out of the flow of income, demand sags, and prices sag with it. On this reading the lost decades were a shortfall of demand, set off by the collapse in asset prices and prolonged by the slow repair of corporate books.

The second explanation looks at production and arrives somewhere else. Fumio Hayashi and Edward Prescott, studying the same decade, concluded that the shortfall was mostly a slowdown in the growth of productivity, how much output the country drew from its workers and machines, alongside a fall in hours as the working week shortened. They argued specifically that the problem was not a broken financial system, since firms large and small could still find financing for the investments worth making. On their reading the constraint was not too little demand but too little productivity growth, and the cure was whatever would raise it.

The two accounts pull in opposite directions, one toward demand and deleveraging, the other toward productivity and supply, and the disagreement has not been settled by thirty years of trying. What can be said is that neither is the whole of it, because beneath both a slower force was gathering. Around the middle of the 1990s Japan’s working-age population stopped growing and began to shrink, and a country with fewer workers each year has fewer hands to produce with and fewer customers to sell to. Aging did not begin the stall, but it deepened the stall and set it, and it is the subject of the next chapter.

The output rose; the wage did not

Step back from the slope of the line and ask what the plateau actually did to the country living on it, and the picture is milder than the word stagnation suggests. Because Japan’s headline growth is output divided by a population that is aging and shrinking, the headline understates how the working part of the economy fared. When Jesús Fernández-Villaverde, Gustavo Ventura and Wen Yao measured Japan’s growth per working-age adult instead of per person, the gap with the United States nearly closed: from 1991 to 2019, output per working-age adult grew about 1.4 percent a year in Japan against about 1.7 in America, and over the two decades from 1998 Japan’s grew a shade faster than America’s. Measured against its own shrinking pool of workers, Japan more or less kept pace with the frontier it had joined. A large part of the lost decades was the arithmetic of a denominator getting smaller.

That leaves the wage, which is a separate question and the one the worker in the opener feels. Output per worker rose; the typical wage did not, which means the additional output went somewhere other than into pay. Where the space between what a country produces and what it pays in wages ends up, in profits, in the returns to capital, in savings, is a question of distribution rather than of growth, and it is a different question from the stall. The point here is only that a flat wage and a rising output can sit side by side, and that neither of them, taken alone, is a country coming apart.

What the plateau was not is a country coming apart. Through the decades of near-zero growth Japan stayed one of the richest, healthiest, and longest-lived societies on earth, with low unemployment. Its people did not grow poorer; they stopped growing richer as quickly. A plateau at high altitude is a genuine loss, the loss of the raises and the churn and the sense of the future arriving that fast growth carries, and that loss is not nothing to a worker whose pay has stood still for twenty years. But a plateau is not a cliff, and the distance between the two is the distance between the fear the chapter opened with and the record.

Where the argument goes next

A stalling country is still a country. It has a central bank, a treasury, a border, and a hundred levers it can reach for, and much of the thirty-year argument over the lost decades is an argument over which levers to pull. What a country does not have is an exit, since a nation cannot pick itself up and move somewhere with better prospects. The next chapter follows the force that did much to deepen Japan’s stall and hold it in place, the aging of its population, and asks whether an aging society is the time bomb it is so often called. Further on, the same logic of decline plays out a scale down, in the regions and towns whose one industry dies and never returns, where the people, unlike the country, can leave, and where the ones who can leave are the first to go.