Search the Site

What Are the Economic Consequences of the Japanese Disaster? A Guest Post by Anil Kashyap and Takeo Hoshi

A local resident holds pictures he found in the tsunami-damaged city of Rikuzentakata, in Iwate prefecture on March 22, 2011. (Photo: NICOLAS ASFOURI/AFP/Getty Images)


From a loss-of-life standpoint, the Japanese earthquake/tsunami may well be at least five times more severe than 9/11. While natural disasters in the past have claimed more lives, it’s extremely rare for a developed country to suffer this kind of catastrophe.
While the economic losses no doubt take a distant back seat to the human suffering, nonetheless there are many important economic questions to be answered. I can’t think of a better pair of people to do so than Anil Kashyap and Takeo Hoshi.
Together, they have been writing about the Japanese economy for 25 years. While they aren’t experts in natural disasters, they have followed the major contemporary events throughout this period, including the spectacular rise in asset prices and the 20-year collapse. (They just finished a consulting report for a Japanese think tank, titled “Why Did Japan Stop Growing?”, which is referenced in their piece below.)
Takeo Hoshi is the Pacific Economic Cooperation Professor in international economic relations at the University of California-San Diego’s School of International Relations and Pacific Studies. Anil Kashyap is a colleague of mine at the University of Chicago; specifically, he’s the Edward Eagle Brown Professor of Economics and Finance at the Booth School. He also co-authored one of the most widely read Freakonomics blog posts ever, “The F.A.Q.’s of Lehman and A.I.G.
Anil and Takeo’s primer on the economic consequences of the Japanese disaster is excellent, and deserves to be as widely read as that earlier article.
The Economic Consequences of the Earthquake in Japan
By Anil Kashyap and Takeo Hoshi

On March 11, an earthquake of magnitude 9.0 hit the Pacific coast side of the Tohoku area in Japan.  Subsequently, a series of aftershocks have continued to hit many parts of Japan.  Even more devastating than the earthquake itself were a series of tsunamis that exceeded 40 feet in some locations and completely wiped out some communities.  The earthquake and tsunami breached the safety mechanisms at nuclear plants in Fukushima and seem to have caused a partial meltdown. The total casualties already exceed 8,000, and thousands of people remain missing.
At this point, it is too early to have a clear idea of what this disaster implies for the Japanese economy and for the rest of the world.  But because we have studied many aspects of the Japanese economy, we have been besieged with questions about the economic effects of this tragedy.  Here is our attempt to come up with preliminary answers for some of the most frequently asked questions.
Q: What is the amount of property loss?
A: While an exact estimate will take months, past natural disasters around the world give us guidelines about the likely effects.  A recent blog by Ilan Noy is a nice summary of what we know from the literature.  The amount of property loss from a natural disaster depends critically on the development stage of the affected country.  To make comparisons across time and space, researchers measure the size of loss relative to the size of the economy (more specifically, as percentage of GDP), rather than the absolute amount in the local currency (yen in this case).  While damage of course depends on the magnitude of the disaster, it is obviously larger in more advanced countries, which have more real assets to lose.  The human casualties, which also reduce the productive capability of a country, are higher for less developed countries, which have relatively poorly reinforced buildings and weak institutions to help cope with the disaster after it occurs.   This combination explains why researchers find an inverse relation between the loss of life and the loss of property after a natural disaster.  Japan, being an advanced country, is expected to have a large amount of property loss and (relative to the size of the earthquake and tsunamis) small human casualties.  Nevertheless, because the quake and tsunami are the largest to strike Japan in at least 150 years, both numbers will be large.
Professors Motohiro Sato and Kazumasa Oguro at Hitotsubashi University guess that the total property loss will be somewhere between 5 percent and 7 percent of GDP (see this Japanese-language article).  As a benchmark, the three prefectures most directly affected (Miyagi, Iwate and Fukushima) account for about 4.0 percent of Japanese GDP and have about 4.2 percent of total population.
Q: How does the loss compare to some past large earthquakes in Japan?
A: In the Kobe Earthquake in 1995, which did not involve a tsunami, the total property damage was estimated to be about 2 percent of GDP.  The death toll was 6,434. (At the time, Kobe also accounted for about 4 percent of Japanese GDP, and that quake was much closer to the city center than this one.)   In the Great Kanto Earthquake of 1923, which devastated the Tokyo area, the total property damage was around 30 percent of GDP.  The death toll is estimated to have been 142,800.  The earthquake that hit Tohoku area on March 11 has already resulted in more deaths than the Kobe earthquake, and the property damage will also be larger.
Q: What is the impact on economic activity from the earthquake?
A: Economists typically separate the loss of life and property, the so-called direct effects, from the follow-on, or indirect, effects that are caused by natural disasters.  These indirect effects are further separated into temporary (or short-run) effects that occur during the first year or two after the disaster, and the longer-run effects.  This distinction reflects the models used in assessing the effects on GDP.  In the short-run, the conventional approach is to add up the various types of spending that can be forecast.  The idea is to forecast the demand for different categories of goods and services to arrive at an estimate for GDP.  Eventually, though, the precision of this approach diminishes because so many intervening considerations come into play.
So for the purposes of estimating long-run effects, economists usually try to focus on the ability of the economy to grow.  These long-run effects are estimated by assuming that growth (broadly speaking) comes from adding capital to the economy, hiring workers, or from productivity improvements.  Thus, longer-run estimates hinge on how the disaster ultimately changes capital accumulation, work force growth, and productivity.
Of course, these two approaches are connected.  The immediate effect of a disaster is to destroy the productive capacity of the economy, so it is possible that some spending cannot occur because the ability to produce the desired goods is temporarily impaired.  Likewise, the growth effects can be slow to manifest themselves if the initial destruction is so large there is inadequate wealth to finance all the desired rebuilding.
Q: What is the immediate effect on productive capacity from this earthquake?
A: Loss of production facilities implies that there will be less capital available to use in production.   The breakdown of supply chains and distributional channels add to the problem.  Currently, an additional issue is the shortage of electricity in Japan, which has already started to affect production.  In addition to these damages to physical capital, the earthquake destroyed human capital, with more than 10,000 lives likely lost.  The people displaced by the earthquake and tsunami may not be able to work and/or have trouble working effectively.
Q: What are the short-run demand impacts?
A: In forecasting the demand side, economists usually look at each major component of the national expenditures (private consumption, fixed business investment, inventory investment, government consumption, public investment, exports, and imports (as negative demand)) and add those up.  Consumption may drop because people now have less real wealth.  They may have to save more to replenish the lost property, which will further reduce consumption.  Fixed business investment is likely to increase as rebuilding starts, although inventory investment will decline at least in the early stage of recovery.  Both government consumption and public investment will rise.  Government consumption rises to support people in the affected areas.  Public investment will rise to rebuild infrastructure in the affected areas.  Exports are likely to fall and imports are likely to increase, reflecting reconstruction demand.  Globally this means increased imports from trading partners to support reconstruction.
Q: What is the overall short-run effect?  Will the earthquake increase or reduce Japan’s GDP in the short-run?
A: As we have seen above, the supply side effect is mostly negative.  The demand side can be positive if the increases in private and public investment as well as government expenditure are larger than the declines in private consumption and net exports.  Thus, the overall short-run effects depend on the strength of the reconstruction demand compared to the size of the lost productive capacity.  The literature finds that GDP often drops following a natural disaster, but the size of the loss critically depends on the developmental stage of the economy.  The loss is much smaller for advanced economies than for developing economies.  Thus, for Japan, we expect that if GDP declines at all, the drop will be very small.
Q: What will happen to Japan’s trade and capital flows?
A: For years, Japan has been selling more abroad than it imports.  The proceeds of these sales have allowed the country to build up considerable foreign assets. The loss of productive capital will make it attractive to rebuild.  Some Japanese company may decide to repatriate foreign assets to finance this rebuilding.   Likewise, the reconstruction demand is likely to increase imports so the trade surplus will likely decline.
Q: What will happen to the yen exchange rate?
A: Forecasting exchanging rates is notoriously difficult.  A classic paper from nearly 30 years ago by Richard Meese and Ken Rogoff showed that a forecast of no change did as well as the leading theories of exchange rate determination.   That said, the conventional theories would identify inflation, interest rates, and repatriation of foreign assets as the main factors driving exchange rate movements.  Because exchange rates involve comparisons of the value of one currency (in this case the yen) relative to another (say the dollar), it is the relative movements in interest rates, inflation, and desired asset holdings that should matter.  Given the many potential movements in these three factors, the standard theory makes no presumption about which way the exchange rate must move following a natural disaster.   But empirically, at least for developed countries, a study by Rodney Ramacharan finds that the exchange rate is likely to depreciate following a natural disaster.
The current Japanese case, however, is perhaps easier to analyze than the general case because inflation and interest rates in Japan have been both very low and unresponsive to economic developments.  Basically prices in Japan have gradually, but steadily fallen for 15 years, and interest rates have been near zero for the entire time period as well.  If one makes the (strong) assumption that these patterns will persist, then a guess about the change in the exchange rate amounts to a guess about whether Japanese owners of foreign assets will want to convert a substantial portion into Japanese assets.  In the process of reconstruction that would be likely, in which case they will sell their foreign bonds (and other assets) and buy Japanese ones.  This would cause the yen to appreciate.
This is indeed what seemed to be already happening until the Bank of Japan and other central banks sold some of their yen-denominated assets to satisfy the increased preference for Japanese assets.  The selling by the central banks (with the implicit promise to sell more if required) seems to have stemmed the appreciation.  Interestingly, the yen also appreciated after the Kobe Earthquake in 1995, even though Japanese interest rates were not quite zero, and the inflation rate then was also positive.  If the demand for repatriation then was strong enough to lead to an appreciation of the currency, despite the potential effects of inflation and interest rates, it is not too surprising that the exchange rate would have appreciated this time.
Q: Will Japan sell the U.S. securities it holds in order to finance reconstruction?
A: Given that a substantial part of the U.S. securities are held by the Bank of Japan, this is unlikely.  As just explained, to prevent yen appreciation the BOJ and other major central banks have already intervened and sold yen assets.  As long as the repatriation pressure remains, Japan is likely to be buying U.S. securities, not selling.
Q: What does this mean for the Japanese government financing?
A: The government needs to spend to restore essential infrastructure in the affected areas.  Also important is the assistance (including housing) for dislocated people so that many of them can go back to work.    This will worsen the Japanese government budget situation, which was unsustainable even before the earthquake.  The government will need a firmer commitment to address its recurring fiscal problems.  If it fails to address the problem in a timely manner, this disaster may lead financial market participants to reassess the terms on which they lend to the government.  Higher borrowing costs would create even more pressure to credibly fix the chronic budget deficits.
There are several steps to stop the budget deficits. Cutting less productive expenditures and concentrating on rebuilding is an obvious first step.   Even before the earthquake, the Japanese government was planning to spend more than twice as much as it collected in tax revenues this year.  It will not be realistically possible to balance the budget just by cutting spending.  So at some point tax revenues will have to increase.  Given the size of the government budget problem, inflation is also likely to be a part of the story to make the government debt sustainable.  The BOJ’s aggressive provision of liquidity is a good start for helping the recovery.  If the BOJ keeps expansionary stance with declined production capacity, the deflation in Japan may finally come to an end.
Q: What is the long-run economic impact of natural disasters?
A: Prior research gives conflicting answers to this question.  Some studies find that a natural disaster depresses the economic activities even in the long-run.  A paper by Eduardo Cavallo, Sebastián Galiani, Ilan Noy and Juan Pantano, however, finds that the negative long-run impact of natural disasters is especially large when the disaster is followed by a drastic political regime change such as the Iranian revolution of 1979.  When they exclude those cases, the long-run economic impact of a natural disaster is almost zero.  Some other studies find that a natural disaster can actually increase the long-run level of output.  This is consistent with the idea of “creative destruction” as an important determinant of economic growth.  A natural disaster, by destroying the poor social and economic institutions that were impediments to economic growth, could enhance economic activity.
Q: What will the long-run impact be on the Japanese economy?
A: Following this line of argument, for Japan, where we do not expect a drastic political system change (although the ruling government might change), we would expect little, or even a very small positive, effect on long-run Japanese growth.  Whether the long-run impact is positive or not depends on what kind of institutional changes this earthquake eventually brings to Japan.  Especially important is what will happen to the major impediments to growth that we identified in our own past work.
In a research report prepared for the National Institute for Research Advancement, we identify two structural impediments for growth in Japan: the zombie problem and bad government regulation.  Following the earthquake, there will be a call for protecting those firms that have suffered from the disaster.  Such protections must be targeted only at those that are affected by the earthquake and otherwise would be profitable.  It would be unwise to assist firms that have been consistently unprofitable.  Likewise, any assistance should be phased out in a timely manner.  Too much and too long-lasting protection after the earthquake can make new zombies.
This mistake happened after the Great Kanto Earthquake of 1923.  To aid the firms that were damaged by the earthquake, the government created a system that essentially allowed companies to get two-year bank loans (that were ultimately financed by the Bank of Japan) that were to be collateralized by their receivables due from the companies affected by the earthquake.  The system was used to help not only the affected companies but also many companies that were in trouble for reasons not related to the earthquake.  The system was to be phased out in two years but was extended for two more years.  During the Diet session on another extension in 1927, the fear that this relief might get terminated (which would lead to many failures) caused a major banking crisis.
In the days immediately following the earthquake, many Japanese noted some government regulations regarding transportation slowed down rescue attempts.  For example, some rescue supplies from foreign governments were delayed because they had to get Japanese government approval to land in Japan.   It was not possible for their helicopters to drop the supplies because they needed to get approval to do so in advance.  Some ships had to wait in the ports until they received  approval to travel Japan’s coastal seas with the same crews that they had used on the open seas.  If Japan realizes the inefficiency that such regulations create in emergency situations, perhaps it will become clear that these same distortions are present in normal times.  Perhaps this could trigger serious, comprehensive deregulation.  If this happens, the current disaster could have a positive long-run impact.


Comments