Japan's Incipient Transformation

Japan appears to be on the cusp of profound structural change: by 2014 its economy will look fundamentally different, and in many ways healthier, than it does today. Not all of the country's problems will be resolved by that point, to be sure, but many will. This conclusion arises from consideration of five points: first, the fundamental problem of excess savings; second, the structural distortions which have ensued from those excess savings; third, consideration of scenarios for how Japan's economy might develop over the coming decade; fourth, identification of the most likely of those scenarios; and fifth, the risks that will still remain.

1) The Fundamental Problem of Excess Savings

The normal pattern for developing countries is that they begin the process of accelerated growth with an inadequate stock of productive capacity. There is accordingly a wide range of opportunities for corporations to invest, the rates of return offered by such investments are good, and households are richly rewarded for supplying capital. Hence both the savings and the investment rates are relatively high. As the economy matures, however, those rates drop because the country now has plenty of plant and equipment and must increasingly devise its own new technologies rather than simply absorbing foreign know-how. As the scope of promising investment opportunities narrows, the corresponding rates of return fall, and households are no longer willing to lend as much to the corporate sector and tend instead to devote more of their earnings to consumption. In general, therefore, a developing country's savings and investment rates start high and then drop over time towards the averages exhibited by more industrialized economies.

Yet Japan never underwent this transition. Here the savings rate remained elevated through the 1990s-well past the point at which the economy had achieved structural maturity. The explanation for this paradox is the disproportionately large share of the population which entered middle age in the 1980s and early 1990s. Since this demographic cohort enjoyed relatively high incomes and was intent upon saving for the exigencies of retirement, its sheer size implied that Japan's household, private-sector, and gross national savings rates would stay much higher than one would otherwise expect of so industrialized an economy.

This penchant for frugality was underscored by the paucity of good investment opportunities available to the public during the 1980s and 1990s. Until the bubble period households earned only very low rates of return on their savings accounts and stock and bond portfolios. They did better from 1985 to 1990, but then came the deflation of the 1990s, which robbed the household sector of a year's gross domestic product (GDP) in wealth and forced people to save even more money in order to achieve their target levels of wealth. Largely as a consequence of these demographic and deflationary trends, the Japanese gross savings rate remained just over 30 % of GDP, as opposed to 20% in the European Union and even less in the admittedly profligate United States, until almost the end of the century.

Contrary to popular belief, this vast quantity of capital represented a grave threat to Japan's prosperity. For excess savings is equivalent to inadequate aggregate demand: unless someone borrows and spends a country's excess funds, they will depress prices and consumption and investment and might thereby push the economy into a prolonged recession or even a depression. Rather than asking why the Japanese economy did so badly during the 1990s, therefore, the more pertinent question is how it managed to perform as well as it did.

2) The Ensuing Structural Distortions

The answers to the foregoing question explain the emergence of some of Japan's most severe structural distortions. The most natural source of the supplementary demand necessary to absorb Japan's surplus savings was the overseas sector. Economic theory states that if a country's savings rate is higher than the appropriate investment rate, the excess funds should be shipped abroad in the form of a current account surplus. There are two rough ways to estimate how big that surplus should in Japan's case have been. First, one can assume that the gross national savings rates of about 20% of GDP which may be seen in most of the other advanced economies is appropriate for a mature Japan and conclude that the excess funds measured about a tenth of GDP. Second, one can compare Japan's 30+ percent savings rate to its probable needs-meaning the total corporate and state demand that would obtain if firms were investing at the same rate as in other mature economies and if the government were running a basically balanced budget. This procedure, too, suggests the volume of excess funds from which Japan suffered was somewhere near a tenth of annual economic output. In other words, the current account surplus should have been about 10% of GDP. Yet it was not. As the second largest economy in the world, Japan could not export that much capital-and, equivalently, that many goods and services-without imposing wrenching change on many other countries. So the long-term trade surplus was rarely allowed to expand much beyond 3% of GDP, and the domestic economy was forced to absorb the rest of the surfeit funds.

In the late 1980s this was accomplished primarily through superfluous corporate investment. Other industrialized countries tend to invest 10%-12% of GDP in plant and equipment, which suffices to support GDP expansion of 2%-3% per annum. This seems to represent a sustainable balance: empirical observation suggests that when the United States, Germany, and other advanced economies get much above that range of capital spending they eventually suffer from bubbles of excess capacity that take years to work through. With this in mind, Japan's behavior looks bizarre. Its ratio of investment to GDP dropped until the early 1980s, as one would have expected of a maturing economy, but then in 1985 that pattern reversed and the ratio subsequently rose to 20% in 1990 and 1991. There were many reasons for this surge in investment, including loose monetary policy, lax financial regulation, and bad banking practices, but its effect was to render firms inefficient and incapable of generating significant profits. Indeed, some calculations of profitability suggest that the non-financial corporate sector was earning less than its cost of capital-and hence destroying investor value-throughout the 1990s.

If the new investment had a devastating microeconomic effect, its macroeconomic impact was paradoxically positive. For the extra capital spending soaked up much of Japan's surfeit capital and enabled the economy to register rapid economic growth in the late 1980s. Indeed, the difference between the actual level of spending on plant and equipment and the appropriate level, estimated again from the experiences of comparable economies, was approaching 10% of GDP in the late 1980s and early 1990s-accounting for almost all of the excess savings. It was thus corporate profligacy that enabled Japan fared so well during the bubble period rather than succumbing to protracted recession as it might otherwise have done.

Beginning in 1992, though, the corporate sector started reducing its investment activity, and the gap between aggregate supply and aggregate demand widened. If unmitigated, this trend might have severely depressed commercial activity. It was therefore fortunate, and perhaps inevitable, that the government budget went into deficit in precisely that year. That deficit, furthermore, grew larger throughout the 1990s in a manner which not only neutralized the contraction in corporate investment but also provided some positive impetus towards GDP growth. Hence the Japanese economy grew in real terms by some 1% per annum rather than shrinking as it might otherwise have done.

Yet dealing with the problem of excess savings through a large current account surplus, a high level of corporate investment, and a steadily increasing budget deficit was not a cost-free proposition, for the consequences of these expedients included great industrial inefficiency, damage to the banking system, and an enormous national debt. Unless one assumes that these structural distortions can continue to grow indefinitely-which seems improbable-then sooner or later Japan must find a more fundamental way to achieve and maintain a macroeconomic balance.

3) Scenarios for Japan's Future Evolution

There are three paths along which Japan might advance over the coming years. The first and most obvious scenario is one in which the savings rate drops by 8%-10% of GDP, thereby obviating the need for supplemental demand and allowing corporations and the government to return to appropriate levels of spending. This could happen either naturally, as those high-saving, middle-aged Japanese retire; or through reforms which force firms to disgorge more earnings to the household sector, thereby helping people reach their financial goals without saving such a high proportion of their income. But the key to this scenario is that some combination of demographics and reform bring down the savings rate.

If one conversely assumes that for economic or diplomatic reasons the savings rate must remain elevated, then the logical answer would be to double or triple the size of the current account surplus-in effect, lending the surplus funds to foreigners so that they may procure Japan's surplus goods and services. But this textbook solution may also prove impractical insofar as the diplomatic and economic realities that constrained the current account surplus in the 1980s and 1990s are still formidable.

This compels us to consider a third scenario, one in which the savings rate remains high and the current account surplus does expand very much. Now the only alternative is a continuation of the status quo, whereby budget deficits and inordinate capital spending soak up all of the surfeit savings. This, too, would enable Japan to maintain economic stability and, perhaps, continue growing at a moderate pace although it would also entail still greater industrial distortions and an even bigger national debt. Once again we arrive at the question of how long this pattern can safely be maintained.

4) Reasons for Cautious Optimism

Why do current conditions give reason for cautious optimism? A quick look at the data on the economic recovery reveals that through the early part of this year foreign demand, especially that from China, provided significant positive momentum. Roughly a quarter of Japan's GDP growth came from an expansion in net exports, while something like another quarter stemmed from corporate investment that was designed to satisfy future overseas demand. Thus Japan has benefited markedly from its massive neighbor's prosperity. But unless one believes that the trade surplus can widen indefinitely without provoking a negative international reaction, sooner or later the recovery will abort unless new sources of demand are found.

More encouraging than the recent recovery, therefore, is a less-remarked decline in the household savings rate. The household capital surplus varied between 6% and 10% of GDP each year through the middle 1990s, then dropped off slightly before plummeting from 1998 onwards. The cause of this sharp decline was demographics. Put simply, a large number of the middle-aged people with high incomes who had previously saved so assiduously were now retiring-meaning that their savings rates fell towards, and in many cases beyond, zero. This trend will almost certainly persist for years to come because so many more workers will be leaving the workforce. There is accordingly little chance that the overall household savings surplus will rise back above perhaps one or two or three percent of GDP. In other words Japan's fundamental problem of excess savings has begun to resolve naturally.

Why has this not yet produced a big change for the better in the economy? Because the corporate sector has largely neutralized the increase in household spending. In order to pay down their bank debt, firms have expanded their savings in the years since 1998, and this has kept the overall private-sector capital surplus roughly as large as it was in the early and middle 1990s. Thus the gap between aggregate supply and aggregate demand has not yet closed significantly.

As a general rule, the beneficial effects of less household frugality cannot manifest until companies have stopped saving so much. This may well happen over the next several years. In the first place, corporate profitability is unlikely to continue at its recent, very high level. A fall in those profits would almost certainly translate into a lower corporate savings rate. In the second place, if corporate governance improves-as the advocates of regulatory and enterprise reform promise-the effect of that more rigorous oversight will be to force businesses to disgorge more of their earnings to households. This, too, will reduce the corporate contribution to private-sector savings while also enabling individuals and families to attain their target levels of wealth without saving as much of their income. It consequently seems likely that firms will eventually join households in moving towards a lower savings rate and hence that the burden of excess capital under which the economy is laboring will decrease.

If the surplus of funds really does diminish, then Japan's output gap will shrink, and then close, and the country will finally be able to redress some of its structural imbalances. Ideally, corporations and the government would avail themselves of the tighter market conditions by curtailing their capital and deficit spending, respectively, as rapidly as possible. This would be a difficult process to manage, since too aggressive a bout of retrenchment might push the economy into recession; but if the spending reductions were implemented very carefully they might eventually bring Japan back into a position where its budget was balanced, or in surplus, and where corporations were investing at international levels of efficiency and generating comparable profits.

Consideration of the political environment, though, suggests this may not occur. The elected authorities and the bureaucrats, except perhaps those in the Ministry of Finance, have strong incentives to stick with a relatively loose fiscal policy in order to keep the economy growing as fast as possible and thereby to gratify the voting public. Corporations, too, have non-commercial reasons to persist in investing above the optimal rate. It follows, therefore, that the best one can expect is a very gradual decline in government and corporate expenditures-less, certainly, than aggregate demand would allow-and hence a continuation of the structural problems in those sectors for some time to come. Yet conversely, this comparatively strong corporate and government demand will underscore the effect of more consumption and ensure that the rate of GDP growth is relatively high in the near and medium terms. In this way a failure to repair its structural problems may well bring Japan more prosperity during the latter half of the present decade.

At a higher level of abstraction this faster growth, relative to the 1990s and early 2000s, will sooner or later close the output gap and transport Japan from one world to another. The country has been stuck in a situation with deflationary tendencies, insufficient demand and falling prices, an eroding financial system, and a compelling need for state stimulus. Sometime after the point at which the output gap closes, however, Japan will enter a realm where demand exceeds supply, inflationary pressures manifest, and the lower savings rate eventually causes the current account surplus to start shrinking. Another important characteristic of this new world will be the importance of supply-side constraints. Over the last 15 years the pace of GDP growth has been dictated by the ebbs and flows of demand for goods and services. Once the slack has been taken out of the system, though, the actual growth rate will fall towards the trend rate of about 1.5% per annum and the only way to improve the economy's performance will be through supply-side reforms such as corporate restructuring and perhaps increased immigration.

5) Remaining Risks

This relatively optimistic scenario is still vulnerable to short-, medium-, and long-term risks. In the immediate future the danger is that aggregate demand might weaken due to deterioration in either consumption or corporate investment; to a "hard landing" in China or a recession in the United States; or to a big, sustained hike in the price of oil. Given the relative sensitivities, the greatest danger is that several of these threats might materialize at the same time, lowering Japan's growth rate and postponing the closure of the output gap. A return to the deflationary circumstances of the 1990s would then be possible. In the medium term, paradoxically, the chief danger is that corporate investment remains excessive, meaning that excess capacity continues to mount, putting negative pressure on prices, employment, and the banking system. For while it is nice in the short term to have corporations spending lots of money, in the medium and long terms corporate demand must be replaced by more vigorous consumption or stronger exports. The ideal trajectory, in other words, would be for a steady but gradual diminution in capital expenditures until Japan arrives at the appropriate level of perhaps 10% of GDP.

Looking further over the temporal horizon, Japan must ultimately deal with its national debt. There has been much discussion of late about whether the gross or the net debt figures give a more accurate picture of the government's financial condition. In the interests of brevity, let's assume that the gross terms are closer to the truth, leaving a more detailed discussion of this issue for the question-and-answer period. On the basis of that assumption, meanwhile, we can take a more detailed look at the government's budgetary prospects.

The Economist Intelligence Unit (EIU) in London has modeled two separate scenarios-a basic case and a reform case-for how the debt is likely to develop over the coming years. The basic case posits a relatively high growth rate in 2004 and 2005, with a decline over the following years to the sustainable rate of about 1.5%; a large increase in the value of assets which the government sells off each year; and an early return to inflation, which causes the real interest rate on the outstanding debt to fall from just under 2% today to 1.4% by 2007. These are generous assumptions, but they still leave the budget deficit widening through the end of the decade and the national debt expanding apace. This trend could lead to an eventual financial crisis sometime in the 2010s or thereafter if investors decide that the government's borrowings are unsustainable.

The assumptions underlying the reform scenario are the same, but now the government additionally tightens fiscal policy. From 2005, for instance, the authorities initiate a series of annual income tax hikes, followed by a hike in the consumption tax from 5% to 7% in 2007. These measures help somewhat; they stabilize the budget deficit; but as you can see from the projections on this chart, they have almost no noticeable affect on the trajectory of the national debt, which continues to accumulate at almost exactly the same speed. The situation would be worse, of course, if markets demanded greater compensation for holding JGBs, for this could push up the real interest rate burden and thereby cause the debt to grow even faster. But even if the bond market remains quiescent, the bottom line is that the fiscal reforms outlined above would represent only the first, modest step in what must inevitably be a long and difficult process of fiscal consolidation. A failure to undertake that consolidation would leave the country vulnerable to an eventual bout of capital flight, with the concomitant political and social dislocation.

Conclusion

In summary, Japan's long slump may well be nearing its end. Greater investment spending and overseas demand have brightened the near-term outlook while a decline in the critically important private-sector savings rate is improving the country's medium-term outlook. Barring a significant shock, or combination of shocks, the output gap should close over the next several years and inflationary tendencies intensify, which will strengthen the banks as well as indebted corporations and households. As the savings rate falls relative to investment, furthermore, the current account will gradually move from a surplus position towards balance and then, perhaps in a decade's time, into deficit. As this happens, foreign institutions will become a bigger source of the money necessary to fund the national debt, and Japan's national finances may consequently become more susceptible to changes in international sentiment. The two major challenges over the longer term, therefore, are managing the national debt and raising the growth rate above its current trend rate by increasing the labor supply, by implementing aggressive structural reforms, or through some combination of both. These remedial measures, however, will depend for their efficacy on the quality of Japan's leadership, with success requiring more vision and decisiveness than Tokyo has demonstrated over most of the last two decades. The good news, though, is that the macroeconomic context in which the government must act will soon be more conducive to reform than at any point since the middle 1980s.

Q: Why do you show corporate investment alongside the government deficit in reference to savings and not alongside government investment?

A: The implication of the elevated savings rate is that there has to be a large government deficit. In terms of aggregate demand-a first-order question, if you will-it does not really matter what form that spending takes. I am not arguing that the government used the money in the right way; the balance between investment, tax cuts and other things was plainly sub-optimal. But on a superficial level, maintaining demand was the most critical objective.

Q: In the longer run, it is a good idea to move the consumption tax, but what is your recommendation for a good tax mix in the short term?

A: I would be hesitant to increase taxes anytime soon. It would be helpful to expand the base of the tax system in order to make it fairer, encourage a better allocation of resources, and facilitate GDP growth, and this could perhaps be done fairly soon. But my inclination would be to leave the overall tax rates where they are for two or three years so that growth continues, the output gap closes, confidence improves, and fiscal retrenchment does not perpetuate the inadequacy of aggregate demand.

Q: Is there a situation where the market gets concerned about the level of debt and yields start to rise? What do you think will collapse first: Japanese bonds or the US dollar?

A: These are fundamentally different subjects. Capital flows are ultimately driven by demographics. All other things being equal, countries with middle-aged populations tend to save more than they invest while younger countries save less and borrow more. Thus you get a natural flow of capital from mature nations to young ones. In this sense the US current account deficit-and even the US government budget deficit-is precisely the same problem as the Japanese savings surplus. The imbalances are two sides of the same coin. It would accordingly be disastrous for Japan, China, and some other countries if the United States unilaterally took steps to balance its current account anytime soon, for then those countries would see their exports collapse and their growth rates fall. In the future, though, as demographics cause Japan's savings rate to decline, global interest rates will rise and the U.S. will be forced, probably gradually, to change its policies. The adjustment will thus stem as much from changes in Japanese domestic behavior as from any alterations in US policy. But the fundamental point is that a lot of these imbalances will tend to correct themselves over time as the demographics of different countries change.

Now the second subject. If it is relatively easy to envisage a gradual correction in the trade and capital flow imbalances, it is a bit more difficult to foresee how the problem of the Japanese national debt will be overcome. If the deficit remains very large for another decade or so, the probability of an eventual crisis would clearly rise. But this could occur with or without an external trigger such as a shock in the exchange markets. It could, in fact, arise from strictly internal Japanese developments.

Q: In the long term, how does Japan deal with the demographic issue as almost every age cohort from age 30 and below is progressively smaller? How does Japan fill that hole?

A: It is erroneous to assume, as some insouciantly do, that the Japanese population will shrink until nobody is left. In two to four decades, the birth rate and population will begin to rise again, eventually enabling the country to increase its trend rate of GDP growth simply by adding labor. The interesting question is what happens between now and then. The decrease in the population is a significant problem but not necessarily overwhelming. If Japan is content with GDP growth of 1.3%-1.5%, for example, then with some redistribution of income and some adjustment in pension programs, living standards could remain reasonably high even if immigration is not encouraged. The trade-off between growth and immigration is thus an internal, political issue. The bigger uncertainty is again the national debt-whether the pace of economic growth will be fast enough, in conjunction with politically feasible tax hikes, to sustain it.

There is also a question of timing. The overall tax rate in Japan is so low that Tokyo could increase its "take" of national income by 10% of GDP, thereby closing today's budget deficit, without reaching much beyond the average tax rate for the EU. Living standards could then still be quite high. But the government cannot safely tighten fiscal policy aggressively before the output gap has closed, and even thereafter must calibrate its tax increases carefully so that they do not push the economy back into recession. Japan must therefore walk a fine line between tightening fiscal policy too rapidly and failing to do so rapidly enough to contain the debt.

Q: Above-trend growth implies that you have a large output gap and, as it closes, you have improvement in public finances. Your budget deficit number does not capture any real benefit from sustained above-trend growth and an accelerating inflation scenario. So how do you fit those together?

A: The EIU estimate of the output gap was somewhere above 3% at the beginning of 2004, which is probably towards the lower end of the range of estimates. The gap was almost certainly much bigger in the middle and late 1990s than it is today, since there has subsequently been a consequential volume of restructuring. But in keeping with our estimate of the output gap, we think Japan could enjoy rapid growth for two or three years, with the actual rate subsequently dropping towards the long-run trend rate. We caught a lot of these dynamics-higher growth in the short term, for instance, and better tax revenues-in that graph. We have also captured the impact of inflation on the average cost of government borrowing. This could in fact be too optimistic inasmuch as prices may not be start rising until 2007 or later. But if we put that caveat aside, yes, inflation helps; it buys Japan a few extra years. But it alone will not suffice to remedy the financial problem.

Q: Why does the deficit not shrink? This inflation must be good for tax revenues.

A: Yes, it is. But what multiplier do you assume? It is hard to come up with a good multiplier on a historical basis because the tax code has changed many times over the last fifteen years. One must therefore make an assumption on the basis of informed guesswork, which we used to calculate an intuitively reasonable multiplier of 1.1. That said, however, the EIU forecasts are subject to all of the usual caveats about modeling exercises. I would therefore not be surprised to see that our deficit projections prove wrong by 1%-2%, and perhaps even a bit more, by 2008. That possibility, however, does not invalidate the main point: namely, that Japan will need to raise taxes dramatically over the next decade or so if it wants to close the deficit and start paying down the debt.

Q: Do you categorize the recent oil price surge as a short-term risk to the Japanese economy and not a long-term risk because it is not a structural or intentional price hike?

A: There are two reasons why the oil shock will probably be more intense in the short term than over a longer time horizon. First, some of the recent price rise is due to speculative pressures that will abate when market conditions clarify. Second, even if there were a permanent change in the price of oil, the initial shock would erode over time as corporations and households adjusted to the new price level by increasing their efficiency. After three or four years the impact on GDP would fall to about zero. Put simply, the medium-term elasticity of demand for oil is much higher than it was 25 years ago.

Q: What is the limit of the external surplus to keep the growth rate high enough to resolve the debt problem? What surplus is acceptable by neighboring countries with increased exports?

A: I have no idea. One can assume that the limit has historically been 3%-3.5% of GDP. I do not see any reason to believe that foreign tolerance has increased much since then, though such a development is conceivable. Another complication is that once the output gap has closed, the extra demand that foreigners might afford is no longer as helpful as in the recent past, for now the constraint is on the supply side. Extra demand from China, the United States, or elsewhere would in these circumstances translate into inflationary pressure and not into sustainably higher GDP growth. The only way to accelerate the growth rate above the trend rate in the long term is to improve the supply-side characteristics of the economy.

Q: Investment in research and development activities to improve the quality of supply may make it possible to export for longer before reaching the limit of supply capability, correct?

A: It is indeed a question of the quality of investment. Japanese corporations presently invest as if the average Japanese worker requires two machines to do what an American does with one. Simply adding a third machine to the assembly lines will not improve that worker's productivity. Overall investment should probably fall to at most a rate of 10-11% of GDP, the level that obtains in the United States, Germany, and other advanced economies that achieve decent GDP growth over long periods of time. But since Japan's trend rate is only about 1.0-1.5%, rather than the 2-3% that characterizes those peer economies, it may be that its efficient long-term rate of capital expenditures should only be 7-9% of GDP. Of course, that remaining investment capital would be spent much more wisely than in the past, and this better quality would raise the trend growth rate slightly. So I would agree with you that the quality of capital spending should improve even as total capital expenditures decrease.

Q: Money has been invested wastefully because the interest the rate is so low here. The financial market gives a signal that you are allowed to waste money.

A: There is something to that argument, but the mistake lies more in the way that firms budget their spending than in the interest rates dictated by the markets. Lets assume that the nominal interest rate which a Japanese company producing for the domestic market might pay on its borrowing is 2%-3% and that measured inflation is currently near zero. Since measurement techniques inevitably overstate the degree of inflation, the actual rate in the economy is probably negative-maybe prices are falling by about 1.0% per annum. This means that the company is borrowing at a nominal 3% and at a real rate of 4%, which is higher than in the United States and many other countries. Hence one cannot reasonably argue that interest rates are too low. The real question is whether corporations are making investment decisions based on the weighted average cost of capital or not. I believe that the evidence suggests that Japanese firms are still not using the right budgeting procedures and that this explains the surfeit investment and capacity as well as the low rate of profitability.

Q: In addition to your scenario, we have the Asian development model of increasing savings first and concentrating on investments in specific sectors for short- and medium-term growth. Also with China's policy, it will have a high current account surplus, meaning a deficit for the U.S. and also Japan. Does this not change your scenario?

A: If China runs a large current account surplus, it may put pressure on the Japanese current account. It is easy, however, to overstate the speed of the adjustment in Japan's external position. Some people have predicted that the current account will go into deficit in 2008. I think that is wrong. If Japan reduces its wasteful investment and its budget deficit significantly over the coming years, the effect of a falling savings rate could be neutralized-by which I mean the savings-investment balance would stay constant-and the current account surplus would not contract for many years. It is likewise important to remember that Japan holds a vast portfolio of overseas assets that are constantly producing income, income which necessarily has an effect on the current account. It will almost take a long time for changes in savings behavior to overcome these realities. My guess, therefore, is that it will be 2012-2016 before the current account goes into deficit

Q: The tax code encourages some firms to report consumption as an intermediate cost to business, so investment is over-reported and the share of consumption too low. So GDP is understated and investment, overstated. This is a potential danger to be aware of.

A: Absolutely. We might add, as you know, that the government is using a flawed method to produce the deflator that it employs to estimate corporate investment, and this similarly has the effect of exaggerating capital spending. I do think, however, that it is safe to say that Japan invests too much and that profitability over the last 15 years has been much too low. These are empirical facts. So at a very high level of abstraction I am comfortable making the assertion that the investment-to-GDP ratio is too high.

Q: You presented the long-term risk of national debt and compared net versus gross data, but would you elaborate a little bit?

A: Broda and Weinstein, who I am told recently presented their analysis at RIETI, say that the net debt figures are more accurate than the gross debt data. In theory net numbers would always be preferable to gross ones just as precise data are better than imprecise data. The difficulty is empirical: are the adjustments to the gross figures that are used to reach the net figures reasonable? This is where I have trouble with the Broda and Weinstein calculations. They arrive at their net estimates by taking the gross national debt of 160%, lowering the estimated value of certain liabilities, and then subtracting the value of various assets to produce a net estimate of 62% of GDP at the beginning of, I believe, 2003. But while they are meticulous in correcting for overstatements of the government's liabilities, they do not scrutinize the various assets that they subtract from the gross numbers. They assume, for instance, that the government's assets can be sold for the values ascribed to them on the government's balance sheets. Is this true? Does anyone believe that Japan could sell the 18 or so percent of GDP it holds in foreign exchange without driving down the value of those holdings? Surely the realizable value of the BOJ's foreign portfolio is a lot lower than its book value. The same is of course true of hard assets: many of them are illiquid and could only be sold for a fraction of their book value. That the government acquired most of those tangible assets before the deflation of the 1990s and early 2000s underscores the probability that they are worth less than what the bureaucrats, and Broda and Weinstein, claim. There are also problems with the reductions that those authors make in the government's liabilities. They say, for example, that Tokyo could net out the debts owed between the central and local governments. This would make economic sense, but it is not politically feasible; for redressing the books in this way would reveal that some of the regional governments are insolvent. To this extent, Tokyo is very unlikely to make the proposed bookkeeping adjustments and, as a result, will remain liable to pay the interest on all of its outstanding debt. Finally. Broda and Weinstein suggest that the government could simply cancel the debt owed to the Bank of Japan; this allows them to subtract that portion of the debt from the government's total liabilities. Reneging on the debt to the BOJ is of course theoretically reasonable, but such an action would in practice frighten private-sector holders of government bonds and could itself precipitate a financial crisis.

What is most remarkable about the Broda and Weinstein paper is that they don't actually recommend making the adjustments they identify. They simply assert that the market understands that the debt is measured incorrectly and hence will never view it as unsustainable. But surely the right thing to do is to test the assumptions underlying the analysis. If Tokyo started selling off its assets and reducing the purported overstatements of its liabilities, we would quickly see which of the Broda and Weinstein adjustments are practically possible and which are not. In this process we would arrive at a reckoning of the national debt that was empirically valid. But until that time their argument should be taken with a grain of salt.

Q: To me it is really the projection of future revenues and costs and the demographics behind that which are the questionable part of the whole thing.

A: Giving a specific example of that, one assumption in the Broda and Weinstein model is that long-term real interest rates stay constant at 2%. This may be realistic if you start with a national debt of 0% of GDP and then borrow another 80-100% of GDP to manage the demographic transition. But if you start with a national debt of 100% of GDP, or 130% or 160%, and then borrow another 80% of GDP, you may quickly discover that the long-term real rate of interest can rise much further than 2%, conceivably triggering serious trouble. In this sense the Broda and Weinstein analysis recalls the old joke about three people stuck on a desert island with only a can of beans to eat. Each proposes a different way to open the can, and when it is the third person's turn-the economist's turn-he says, "assume that we have a can opener". If the islanders had a can opener, of course, they would not have had a problem opening the can in the first place. By the same token, if we assume that interest rates cannot rise in response to external or internal shocks, the odds of a debt crisis are artificially and unrealistically reduced. Sadly, assuming away the fundamental risk in this way is not something the government can do.