Paying for It: Part 2 - The Great Kanto Earthquake

Think coronavirus is bad?

The catastrophic Great Kanto Earthquake struck on 1 September 1923, leveling most of Tokyo and Yokohama. The casualty statistics are shocking: 91,000 killed, mostly in Tokyo and Yokohama; 13,000 missing; 52,000 injured; 69,000 houses lost or damaged. At the time it was estimated that 3.4 million people were directly effected – approximately 6% of the population.

The damage to national wealth was put at Y5,274 million, approximately 4% of total estimated national wealth of 1924, and 44% of 1923 GDP. Of that total, Y2,136 million worth of merchandise was lost and Y1,874 million worth of buildings, equivalent to 8.2% of the total building-stock. Perhaps we should be skeptical of these numbers: the Y2,136 million claimed as loss of merchandise and inventory seems an astonishingly high number (particularly when compared with the Y869 million lost in household goods), and suggests there was no shortage of people willing to inflate their losses in response to government relief actions.

This estimate of losses does not, however, include the impact on the stockmarket. In the event, however, these seem to have been surprisingly light. The bond market reopened for business on October 16 (in IBJ’s banking hall), spot transactions started again in the ruins of the exchange on October 27, and futures trading restarted on November 15. But, buoyed by a Y13 million rehabilitation loan from IBJ, when TSE proper reopened on November 15, its shares, which had been quoted at Y121.9 on August 30, opened at Y107 – a loss of only 12%.

Responses

The government’s bureaucratic responses, and its immediate provision of liquidity were as one might expect. The government provided relief financing and, cut duties on the import of reconstruction materials. Not surprisingly, this resulted in a flood of speculative imports, which in turn blew out the trade deficit in 1924 to record proportions (6.7% of GDP), which in turn cut the gold reserves, drove the yen down from around 49 at beginning of 1923 to 38.5 by October. Such negative developments (from the point of view of a political leadership unwilling to abandon its gold-standard aspirations) constrained the government’s role in reconstruction.

More significant, in the longer term, were the government’s arrangements to contain the immediate economic and financial damage. On Sept 7, the government declared a 30 days moratorium on all obligations contracted before Sept 1 and payable during September, for all debtors having their residence or place of business in the stricken area. In addition, when that moratorium ran out, BOJ was empowered to rediscount all bills discounted by banks and maturing before Sept 1925, which were covered by the moratorium, with the government offering to indemnify BOJ for losses on these transactions up to Y100 million.

In other words, when the deadline for payment of a note issued or to be paid in the earthquake region came around, the note could be taken into a bank to be discounted. The bank would subsequently take the note to BOJ, which would rediscount the note stamping it with the words “Earthquake Bill.” In the event that the debt turned out to be bad, or simply not paid on time, the government would guarantee BOJ up to Y100 million.

Many companies were saved by this procedure, but at a cost to the financial system which was only revealed over time.

First, this de facto BOJ guarantee for bad debts allowed banks – who were still reeling from the implosion of the 1920 bubble - to avoid making credit judgments they didn’t want to make, and generally encouraged fraud. Second, it would also lead to a rise in bank loan/deposit ratios and, conversely, a rise in corporate leverage ratios. At the end of WW1 banks’ LDR stood at 94.2%: by 1923 it stood at 115.7% - a level it would not reach again until after WW2. The aftermath of the earthquake provided banks with a final chance to re-leverage, courtesy of the government’s guarantees. They would soon (in 1927) discover that those government guarantees did not solve the underlying problems of their portfolios. For the next 17 years, banks basically built this ratio down – in a way which was bound to be deflationary.

As for corporate leverage, 1923 marked the start of a build-up in leverage ratios which continued throughout the 1930s (mainly reflecting the rise in debt securities markets). Releveraging in a deflationary environment is asking for trouble.

Third, and crucially, the technique of rolling debts, adding liquidity, and giving a guarantee that would deal with only part of the problem, didn’t work. The government had indemnified BOJ for losses on these bills to the tune of Y100 million, but this clearly was insufficient to deal with all the bad debts left behind by the earthquake. The amount of “earthquake bills” has been estimated at Y2,100 million, equivalent to 18.7% of all bank loans. Frankly, I don’t know how this estimate was arrived at – most probably it follows the “lost merchandise” figure, about which we have already been skeptical.

In practice it came to less than that. The face value of the bills payable at the end of March 1924 and rediscounted by BOJ came to Y431 million. That’s only about 4% of bank lending, but it was also 33.5% of all industrial bonds outstanding. By November 1924, the unpaid amount of “earthquake bills” had fallen to Y276 million (21% of bonds outstanding), but their liquidation was slow and even at the end of 1926, there were still at least Y207 million outstanding of these identifiably bad loans circulating in the system, equivalent to 11.2% of all outstanding corporate bonds.

These earthquake bills stressed the financial system in at least three ways. First, they cluttered up BOJ’s balance sheet, leaving it less able to lend to other parts of the financial system. Secondly, this amount of bad debt undermined the overall health of the money market, thus muting any impact BOJ could have made still further. Third, the earthquake bills also undermined banks’ balance sheets: because they had to depend on foot-dragging political decisions of the government with regard to loans, banks holding these bills were gravely weakened.

The banking industry was in no fit state to cope with these pressures. In the aftermath of both the implosion of the Bubble in 1920, many banks were already on shaky ground, with the industry consolidating fast. Japan had lots of banks, mainly extremely small, extremely local and inevitably horribly entwined with local business concerns.

“Companies . . . commonly went to great length to cook the books so that their assets and liabilities tables showed a profit even when the company was awash in red ink. Firms would fail to write off unrecoverable loans and accounts receivable and would appreciate the book value of securities or real estate in their possession, even when the reverse was the case, so that it would appear that the firm’s assets had increased and that it was making a profit. There were many instances of this window dressing, complete with bogus dividends, for to report a loss would blemish the company’s credit and cast doubt on the management skills of the directors.” “Lectures on Modern Japanese Economic History, 1926-94” Nakamura Takafusa.

Essentially, the earthquake bills mess added to the problems of a financial system already carrying unacknowledged losses from the post-war bust but which was already actually encouraging corporate re-leveraging via the bond market despite the underlying unremitting deflation. All this came tumbling down in the financial crisis of 1927.

Now of course, history does not repeat itself. But I think the lesson to be drawn from rescue packages for the Great Kanto Earthquake disaster are these: if a government believes in debt write-offs or forgiveness, and tries to engineer this through possibly unfunded central bank largesse, do not be surprised if that largesse is abused, and in the aftermath, the strain of the solution examines and intensifies fractures and vulnerabilities already concealed in the financial system.

Which economies, if any, currently suffering a coronavirus economic coronary, might look like that?