Presentation by Carlo Cottarelli, Director, Fiscal Affairs Department, International Monetary Fund
At the London School of Economics and Political Science (LSE)
Let me start by thanking LSE for inviting me here today. I am here to present the September 2011 issue of the Fiscal Monitor (Slide 2) but I would also like to take this as an opportunity to discuss some topics that were discussed in previous issues of the Fiscal Monitor, the new IMF publication that we started issuing in 2010.
I will discuss in turn (Slide 3):
- the state of public finances in advanced and emerging economies before the 2008 crisis
- the effect of the crisis on the fiscal account
- the risks arising from the current situation, particularly for advanced economies
- and our current fiscal policy advice to address those risks.
So let me start from the condition of public accounts before the crisis.
In a nutshell the state of fiscal accounts had never been so weak in decades in advanced countries.
The data that I am going to show you come from a public debt data base that we made available last year on our website: it reports public debt data for over 150 countries, with some time series starting two hundred years ago. But here I will focus just on the last sixty years.
Let’s start with the G7 (Slide 5). This is their average public debt to GDP ratio. It is clear that, starting with the early 1970s the debt ratio had been on an upward trend and that in 2007 it had already reached the level reached in the aftermath of the Second World War. But this underestimates the weakness of the fiscal accounts. In the aftermath of the Second World War, advanced countries still had a relatively young population. In 2007, population aging and its impact on the fiscal accounts was clearly a problem, with the net present value of pension and health care spending increases (an implicit debt to be added to explicit debt) amounting to hundreds of percentage points of GDP. In addition, financial markets in the early 1950s were much less complicated and much more constrained than they are now with banks often obliged to purchase government paper from their own country and limited investment abroad. It was then much easier to roll over public debt than it is now.
Within the G7, debt in 2007 was high in all countries, with the exception of Canada and the United Kingdom. The latter on account of the large size of its financial sector with respect to GDP was facing large contingent liabilities in case of a financial sector shock, which unfortunately materialized.
Things were somewhat more mixed in some smaller advanced countries, but not in all (Slide 6). Debt ratios were low outside Europe, like in Korea, Australia or New Zealand, but were high, historically, in several countries, like Portugal and Greece.
This high level of public debt was seen by some as a drag on long-term growth by many, but most economists, and definitely financial markets based on the spreads prevailing at that time, thought that advanced countries could not be affected by open fiscal crisis, by a roll over crisis. Following events will show that this was not true.
What was the effect of the crisis on the fiscal accounts of advanced economies?
So let’s start with advanced countries focusing on what happened in 2008-11. I am going to show you a fiscal movie which describes the evolution of deficits and debt with respect to their level in 2007 (Slide 8-13). So, on the horizontal axis you will see the increase in debt with respect to 2007 and on the vertical axis you have the increase in debt with respect to 2007. At the start, all countries are at the origin and we will see how things evolved over time. The size of each balloon is proportional to a country’s GDP. So for example, this tells you that already in 2008 the debt ratio in the U.S. had increased by 10 percentage points over the previous year and the deficit ratio by about 4 percentage points. There was already a sizable deterioration in most countries but the big shock was in 2009 with major increases in the deficits and debt. Nothing much happens in 2010 to deficits: deficits have edged down mostly as a result of the improvement in economic conditions, but debt keeps rising fast. By 2010 debt has risen to 25-30 percentage points above the 2007 level for several countries.
In 2011 you start seeing some more considerable fiscal adjustment, including in the United States. It is interesting to note that fiscal policy in the U.S. turned out to be considerably tighter than expected: the Obama administration, with the stimulus approved in December, intended to have a moderate fiscal expansion in 2011. This did not materialized partly because revenues arising from financial market profit and asset price increases in 2010 were larger than expected, partly because of the difficulty in raising spending due, first, to the absence of an approved budget, the possible government shut-down, and later to the debt ceiling saga.
Policies have now already taken shape for 2012, but there is a lot of uncertainty about U.S. fiscal policy due to the stalemate in Congress. This is our best projection of what would happen across countries at unchanged policies. The decline in the deficit in the United Sates is remarkable—more than 2 ½ percentage points of GDP—and I would say too fast given the risks ahead for growth. If the American Job Act, the support plan, sent to Congress by Obama were approved fully, this is what would happen: essentially no fiscal tightening: the fiscal tightening would be postponed to next year, when, hopefully, private sector demand would be stronger. But this short-term support is facing a lot of opposition in Congress. A lot of opposition is also likely to be found by the medium-term plan put forward by Obama a couple of weeks ago—called Living Within Our Means and Investing in the Future—a plan that is meant to ensure that confidence in the U.S. public finances would remain strong in spite of the short-run stimulus. In sum, due to the division in the U.S. Congress, the U.S. has not yet clarified either its short, or its long-term policy course.
Note that the output shocks affected the fiscal accounts both on account of the automatic revenue loss and expenditure increases related to a recession, and on account of discretionary measures that were undertaken. A lot of focus has been put on whether it was appropriate or not to have discretionary stimulus in 2009-10. The fact is that, as illustrated here (Slide 14), most of the increase in the debt to GDP ratio that we are experiencing in advanced countries comes from the automatic revenue loss, with only about 20 percent coming from discretionary fiscal stimulus.
So, the fiscal shock has been huge. A good metric to see this is to compute how long it would take to reverse the increase in public debt. We have updated in this Fiscal Monitor a calculation that we have presented earlier, looking at how much countries would have to improve their cyclically-adjusted primary balance in order to lower the debt to GDP ratio to 60 percent by 2030, assuming that they want to improve the primary balance within the next 10 years and then to keep it constant at that level for another 10 years. This is the starting point in 2010 for the primary balance, the primary cyclically adjusted balance, and the overall balance for the average of the advanced countries (Slide 15). Assuming the target is to lower the debt ratio to 60 percent (we use 80 percent for net debt for Japan, where debt is much higher) the primary balance would have to improve in cyclically adjusted terms by about 8 percentage points of GDP to reach about 4 percentage points of GDP. The improvement in the headline primary balance is even larger. And advanced countries would have to achieve this when pressures from health care spending and pension spending would be rising by about 4 percentage points of GDP due to demographics and other factors. It is definitely a daunting task.
This chart (Slide 16) compares, country by country, the level of the primary balance that would be needed in this exercise against the level of the primary balance ever reached in any ten-year period before. It is clear that for about half of the countries lowering the debt ratio to below sixty percent would require them to do something that had never been done before.
So, what are the risks now? I will focus the rest of my presentation on advanced countries and, specifically on the euro area, the United States, and Japan.
A good place where to start to assess risks is by looking at what markets believe, that is by looking at interest rates at which countries are borrowing.
Let me start with a simple chart (Slide 18). This plots for all euro area members the deficit to GDP ratio in the vertical axis against the debt to GDP ratio. The size of each balloon is proportional to the yield differential between the country in question and Germany that is here regarded as the risk free asset.
It is visually impressive that all countries that have large spreads (Greece, Portugal, Ireland, Italy, Spain, Cyprus and Belgium) lie on the right hand side of this line, which I would call therefore “safety frontier.” So I am going to put forward a very “scientific” theorem: a sufficient condition to be safe is to be within the frontier. A necessary condition to be in trouble is to be outside the frontier. So, basic fiscal fundamentals do matter.
But let’s go back to the 2011 situation. Being in the danger zone is a necessary condition to be in trouble: but it is not a sufficient condition. You have France, but here one could argue that interest rates have indeed been on the rise in France. More significant are the cases of the United States and Japan, which are positioned in the neighborhood of the Irish planet.
Why are they not in trouble?
Let’s consider some explanations, which I believe will shed some light on the policy options that countries are facing. I will consider in turn: past and future fiscal trends; pressures from health care and pension spending; banks; growth; and sources of financing.
First, hypothesis: this figure (Slide 20) is static: it only focuses on the current deficit and debt but not on past and future trends. So let’s look at these trends.
This is the deficit in 2009 and in 2011 (Slide 22): a large improvement in almost all countries, including in the United States, but not in Japan.
Note also that the decline in the deficit in Europe was larger than anticipated. This is the decline that we were projecting two years ago, and this is the actual decline (Slide 22): there were good in Europe, less so in the U.S. and no good news in Japan.
Let’s look ahead. Here we plot the increase in the debt to GDP ratio for Japan and United States during 2012-14 against that of the euro area countries (Slide 23). It does not look good. Note also that the projected increase in the debt ratio for several euro area countries is actually smaller than we were projecting, say, two years ago. So, there has been good news for these countries: no good news for Japan and the United States.
So, altogether it does not seem that recent or future fiscal trends explain why U.S. and Japan are not under market pressure.
Second explanation: Longer-term trends in pension spending and health care spending. How do euro area countries compare with the United States and Japan? This is shown here (Slide 25), which focuses on the net present value of future pension and health care spending across countries. As you can see, Japan is not in a bad shape but the United States is in a terrible shape, owing primarily to health care spending.
So, having ruled out what is not important, let me tell you what I believe is important.
First, the condition of banks. I do not want here to reopen the issue of the need for stronger capital buffers in European banks. But it is enough to look at bank stock prices to understand that there is concern about the state of European banks, while there is much less concern at present about U.S. and Japanese banks. This said, there is an element of circularity here, as a good part of the weakness of European banks arises from the holdings of government papers whose yield has increased, so, in itself, this cannot explain the increase in yields.
Let’s keep looking.
Growth: this chart (Slide 27) plots the interest rate spread against the growth rate in 2010-11. There is definitely a relationship, although it is an explanation that holds more for the United States than for Japan, a low growth country. Moreover, the correlation is much less clear if we take a longer time period as it is done here (Slide 28). Finally, one question is to what extent the increased level of public debt will be a burden for growth in the U.S. in the future. There have been a few studies recently by Rogoff and Reinhart, Kumar and Woo, and Cecchetti providing econometric evidence of the fact that, at least beyond a certain threshold, high debt is bad for growth (Slide 29). The United States is now well beyond that threshold, which is found to be at about 80, 90 percent, although it is lower in other studies. So, the question is whether in the future the United States will continue to maintain its comparative edge vis-à-vis Europe in terms of growth.
Let me move now to what I believe is the most obvious explanation for the differences that we observed in yields across countries: their investor base.
This chart (Slide 31) looks at the holding of government paper by various classes of investors. The most apparent thing is the very limited reliance on foreign private sector financing in the liabilities of both the United States and Japan. In Japan almost all debt is held by domestic residents, mostly financial institutions. In the United States, over one third is held by foreign central banks: the so-called exorbitant privilege of the United States, the fact that short-term dollar assets are held as reserve by central banks. Another 20 percent is held by the FED. The share of U.S. debt held by foreign banks has been rising since 2000, reflecting mostly increasing holdings by China (Slide 32).
In contrast the share of foreign financing for the euro countries under pressure is much higher, although it is overestimated in this picture by including ECB purchases as nonresident holdings. One can note that the share of foreign holdings is correlated to the interest rate spread for the countries with weaker fundamentals (Slide 33).
Putting all this together, we tried to assess in a more consistent way why CDS spreads differ across countries in 2011 through some simple cross-country econometric regressions (Slide 34). It is worth stressing some key results (see details in Annex I). They illustrate the current short-termism of markets, which makes fiscal policy management more difficult. First, fiscal variables are important, but markets are now focusing primarily on short-term developments (the projected primary deficit and debt in 2011). Future variables (future debt and long-term spending trends) do not seem to matter much. Second, projected growth is important (higher growth leading to lower spreads), but, again, short-term growth is what matters, rather than potential growth. One unpleasant implication of this focus on short-term output growth is that, if the fiscal multiplier is sufficiently large (higher than 1.2-1.3 based on the estimated coefficients), a fiscal tightening can lead to a rise in spreads: the improvement in the deficit tends to lower spreads, but the short-term decline in GDP, acting also through the short-term rise in the debt ratio, tends to push spreads up. Third, central bank financing (either from your own central bank or from other central banks) is important in lowering interest rates, as long as it is not inflationary.
I draw several policy conclusions from all the above discussion. Let me start from the United States and Japan.
Their fiscal accounts are very weak and these countries do not yet have a clear plan to stabilize rapidly and then lower their debt to GDP ratio: discussions on a grand bargaining for the United Sates are continuing in Congress. As to Japan, there is a plan but it is not very ambitious, involving the stabilization of the debt ratio only in 10 years or so. Their strength is primarily their financing structure. The United States remains exposed to changes in the decisions of other central banks. Japan is exposed to changes in the attitude of its bank and other financial institutions towards what is profitable. Altogether, these countries need to quickly clarify their policy intentions.
What are the lessons for European countries under pressure?
First, they need to adjust their fiscal fundamentals. The adjustment is taking place, though. This is how these countries would move over the next two years (Slide 37).
Things would be improving but many would still be in the danger zone, which means that interest rates will likely remain high, hopefully not as high as they are now. Italy would be here: we project a deficit of 1 percent of GDP higher than the authorities’ target of balanced budget primarily because we have lower growth. One part of the package—the reduction of tax expenditure amounting to some 20 billion in 2013—has not been yet identified in specific terms—the government will have to do this— but there is a safeguard clause implying that tax expenditures will be cut across the board if no specific decisions are undertaken.
So fiscal adjustment is coming, but it is not enough. I would underscore three more points.
First, growth. Structural reforms are needed to boost the growth process. Higher growth has a huge impact on the fiscal accounts over the medium term. For countries like the European one with a revenue ratio of at least 40 percent, raising the GDP growth rates by one percentage point would lower the debt to GDP ratio by 30 percentage points after ten years, if all additional revenues are saved (Slide 39). However, short-term growth will suffer from the fiscal tightening if not offset through adequate financing (see below).
Second, there is a need to increase the capital buffers in the banking system: more capital will support confidence and will allow an increase in credit as needed to support the growth process.
Finally financing. In an environment in which markets are focusing on the short term, there is a risk that the necessary fiscal tightening leads to a rise in spreads, if not offset by the provision of financial support to the countries that are adjusting. We have seen how important adequate nonmarket financing is for the United States. Of course, I am not suggesting that these countries become reserve currency countries. But, in the short run, the support of the ECB will continue and, for the program countries, financing from the EFSF and the IMF remains available, which in turns depends on the implementation of adequate reforms. The July 21 agreement within the euro area to support as long as it takes countries that are adjusting is critical in this respect. I think that, ultimately, the political will to do whatever it takes to support countries in trouble that are adjusting and to support the euro area is there. But it needs to be communicated without hesitation. And unfortunately, communication has not been very good so far.
I will stop here.
Thank you very much.
Annex I – CDS spreads in advanced countries
By Carlo Cottarelli and Laura Jaramillo1
The attached table presents some OLS cross-section regressions of 5-year CDS spreads for 31 advanced countries (Slide 34 of the accompanying PowerPoint presentation).2 The dependent variable is the log of CDS spreads (average for the first three quarters of 2011). The regressors are:
pb2011: the general government primary balance over GDP ratio in 2011 (as projected in the Fall 2011 Fiscal Monitor of the IMF), but only for countries that are members of the euro area (the primary balance for other countries is not significant and when the primary balance for all countries is included the magnitude of the coefficient drops by half)
gdebt2011: the general government debt-to-GDP ratio at end 2011 (also as projected in the Fiscal Monitor). For Canada and Japan, net debt to GDP is used.
dfcb: the general government debt of the country in question held by its national central bank as a ratio to GDP and, in the case of Japan, the U.K. and the U.S., by foreign central banks, based on the latest available data.
infl2011weo: the inflation rate projected in the World Economic Outlook of the IMF (also Fall 2011 issue) for 2011
rgdp2011: the real GDP growth rate for 2011 from the same source
pen2010: the net present value of the increase in public pension spending during 2010-50 as a ratio to GDP (as projected in the IMF Fiscal Monitor)
health2010: the net present value of the increase in public health care spending during 2010-50 as a ratio to GDP (as projected in the IMF Fiscal Monitor)
pb2014: the primary balance to GDP ratio for euro area countries in 2014 (as projected in the IMF Fiscal Monitor)
potlr: projected potential real GDP growth, averaged over 2011-16
The first column of the table reports a general specification in which all variables are included. The following columns illustrate the specification search, with insignificant variables being dropped one by one.
The regressions illustrate the focus of financial markets on short-term variables. Neither the future primary balance, nor future spending trends for pension and health care, are significant (future debt ratios were also insignificant). Future potential growth enters the regressions with a sign that could not be justified from an economic perspective (higher potential growth leading to higher spreads) and a relatively low t-statistic.3 It was therefore dropped during the specification search.
Based on the regression in the last column, CDS spreads in 2011 appeared to be affected by:
• the fiscal position in 2011, as described by primary balance and debt ratios: the coefficients are broadly in line with what has been found by previous econometric work. Given the specification in logs for the CDS spread, they imply that the larger the initial level of the CDS spread, the larger the impact on spreads, in basis points, of an increase in deficit and debt ratios; consistently, a weakening of fiscal variables has a more negative impact in countries with higher initial deficit and debt ratios.
• purchases of government paper by central banks: a higher coefficient than the one on the debt ratio implies that the effect of purchases by central banks (and by foreign central banks for reserve currencies) goes beyond the effect of reducing the overall supply of government bonds sold to the market. This probably reflects confidence effects provided by the presence of the central bank in the market. Note, however, that this variable does not include purchases by the ECB. When these were included separately they appeared with the “wrong” sign and when added to dfcb the significance of the variable declined. This may be due to measurement problems, as the ECB does not publish its purchases country-by-country so market estimates were used. It may also mean that purchases that are not conducted transparently (at the country level) are not very effective. Finally, the wrong sign may reflect simultaneity problems (the ECB intervening on countries whose spreads are already high).
• inflation: this variable is highly significant, and thus implies that central bank purchases are effective in moderating spreads only if they are not inflationary. Given the high demand for liquidity in 2011, this was probably the case for many countries.
• real GDP growth: one should expect GDP to enter a spread equation as high growth is good for fiscal variables (affecting directly the dynamics of the debt-to-GDP ratio and making it easy to run primary surpluses, certainly in the short run and possibly over the longer term too). However, what should affect markets’ perceptions of fiscal sustainability is long-term growth, not short-term growth. The fact that markets are focusing in 2011 on short-term growth developments may reflect strong risk aversion after four years of market turmoil. The unpleasant implication of this short-termism is that a tightening of fiscal policy may raise rather than reduce spreads if it is accompanied by a decline of GDP (with respect to the baseline). Indeed, the estimated coefficients imply that this would happen for a fiscal multiplier higher than 1.2-1.3 (in this case the primary balance would improve, but the debt to GDP ratio and the CDS spreads would increase).
These results are preliminary. One important issue is to understand the extent to which these results reflect the particular state of markets in 2011, rather than more permanent features, something that a cross section cannot shed light on. Moreover, it would be important to assess the direct effect on spreads of other variables (such as exposure to contingent liabilities from the banking sector). Potential simultaneity issues (e.g., between spreads and growth) also deserve additional attention.
1 Raquel Gomez Sirera provided excellent research assistance.
2 The country sample includes Australia, Austria, Belgium, Canada, Cyprus, Czech Republic, Denmark, Estonia, Finland, France, Germany, Greece, Hong Kong SAR, Iceland, Ireland, Israel, Italy, Japan, Korea, Malta, Netherlands, New Zealand, Norway, Portugal, Slovak Republic, Slovenia, Spain, Sweden, Switzerland, United Kingdom, and United States.
3 Past potential growth was also included but presented the same problem.