June 19, 1993

A Japanese puzzle
JAPAN: Economic Focus - Japan's Capital Flows

Economists have had little success in explaining one of the most striking economic facts of the 1990s: the ebb and flow of Japanese capital. A new approach offers an answer, and may shed light on an even bigger question students of economics are taught that the current and capital accounts of the balance of payments are tied together by an identity: suitably defined, the current-account balance is equal and opposite to the capital-account balance. In other words, if a country produces more goods and services than it consumes (ie, if it has a current-account surplus), it must lend the proceeds to the rest of the world (ie, it will have an equal and offsetting balance on its capital account). A common mistake, however, is to assume that the components of the capital account move in step. For instance, a current-account surplus need not go hand in hand with an outflow of both long- and short-term capital. Japan in the late 1980s and 1990s offers evidence. As the chart shows, from the middle of the 1980s Japan's outflow of long-term capital often greatly exceeded its current-account surplus. That is, despite its current-account surplus, Japan was borrowing short-term money from abroad in order to invest far larger sums in overseas assets. Then, after 1990, this peculiar relationship went into reverse. Despite a growing current-account surplus, long-term capital flowed into Japan, implying a correspondingly huge outflow of short-term capital. Traditional economic theories find these shifts hard to explain. As a result, one of the most striking economic phenomena of the late 1980s and early 1990s has been one of the least well-understood. A new study by Richard Werner, an economist at Oxford university and a visiting researcher at the Bank of Japan, sets out to change that. If Mr. Werner is correct, the implications are of great interest to students of other economies, too. Start with some familiar facts. It is well known that in Japan in the late 1980s, the creation of domestic credit exploded, and that this monetary expansion fuelled a steep rise in the prices of domestic assets (especially land). At the beginning of the 1990s, this cycle went into reverse: the supply of credit was sharply squeezed, and the price of land and equities collapsed. At first sight, this credit cycle seems intimately linked to the pattern of capital outflows. And so it is, argues Mr. Werner. But understanding the connection calls for a rethink of central aspects of orthodox monetary theory. According to conventional models, a monetary expansion such as Japan's in the late 1980s ought to have depressed the yen, driven the country's current account into deficit - ie, greatly reduced its capital outflows - and sent inflation soaring. None of that happened: the yen strengthened; although it narrowed, the current-account surplus remained; capital outflows expanded hugely; and inflation hardly budged. It seems obvious (especially to those who work in financial markets) that the credit cycle was somehow linked to the shifting pattern of long-term capital flows. But how is that to be reconciled with what is known about the way money affects economies? Mr. Werner's answer is that conventional monetary theory is wrong. The traditional approach starts from the "quantity theory", another accounting identity, which posits a link between the rate of increase of money and the rate of inflation in an economy. For years this relationship seemed stable enough to rely on for policy purposes, but in many countries it broke down in the 1980s - hence, in Japan, the combination of exceptionally rapid monetary growth and very low inflation.

Money is roughly the same as credit, thanks to (you guessed) another identity. A bank's balance sheet tells you that bank deposits (broad money) must be about equal to bank lending. Mr. Werner argues that money can therefore be disaggregated in a very informative way. Divide it into loans that finance purchases of goods and services, and loans that finance purchases of assets; then compare those measures of money with corresponding measures of inflation. He does this for Japan, building a two-part model. The first part aims to explain inflation in the prices of goods and services, the second part inflation in the prices of assets. The model seems to work, passing the obligatory array of statistical tests with flying colours. Interestingly, though, by attacking the conventional theory, Mr. Werner seems to have resurrected it. The goods-and-services side of his model shows a stable relationship between Japanese money (ie, credit for "real" transactions) and consumer-price inflation throughout the turbulent 1980s. And the financial-assets side of the model accounts for inflation in land and equities.

Where we came in
This, in turn, accounts for capital outflows during the years when credit was expanding, and inflows when credit was squeezed. How? In effect, the soaring value of land provided the collateral against which Japanese firms could borrow at home, to buy assets abroad. Since Japan combined this asset-price inflation with stable goods prices and an appreciating currency, it achieved a possibly unprecedented feat: through reckless creation of "financial" credit, it printed the means to buy real assets from foreigners. At first sight, Mr. Werner's approach also seems to make sense of Britain's experience during the 1980s. That is another case which has perplexed many species of monetarists, and where a disjunction between the real and financial economies has seemed to play a crucial role. Mr. Werner has proposed a model which is monetarist in spirit, but which sheds light on events that conventional monetarist theory has been at a loss to explain. Expect to hear more about it.

"Towards a Quantity Theorem of Disaggregated Credit and International Capital Flows with Evidence from Japan." Paper presented at the annual conference of the Royal Economic Society, April 1993.