Why Are Long-Term US
Yields Low?

Since June 2004 the Fed Funds target rate has increased significantly, but nonetheless the 10-year yield was virtually unchanged at the end of November 2005, compared to June 2004. Long-term yields appear low in terms of both expectations of nominal growth and traditional empirical models. On a global scale too, long-term yields are low in relation to the cyclical position of the global economy. The article reviews a number of explanations for the low long-term yields, as discussed in the media and academic literature. The starting point is US yields, which can be perceived as the hub of the global formation of interest rates.

The decline in nominal yields over the last 15-20 years is primarily attributable to lower inflation expectations as a consequence of lower realised inflation. In addition, the accompanying stability of inflation rates has diminished uncertainty of future changes in inflation. This has reduced the inflation risk premium and contributed to lower interest rates.

The economic development in Asia entails that global production capacity is now expanding, and will continue to do so, which tends to dampen inflation. This also contributes to dampening inflation expectations and interest rates.

Since 2003 the global cyclical position has been favourable, but investments have not increased to the expected extent. The dampened demand for real capital has helped to bring real interest rates down.

Finally, in recent years pension companies and Asian central banks have demanded long-term bonds in particular. This has contributed to the low yield level. The high demand for bonds is partly derived from other considerations than real rates of return, namely in the case of pension companies the hedging of liabilities, and in that of Asian central banks the fixed-exchange-rate policy.

The overall picture is that long-term yields are low, but not inexplicably low. The sustainability of the long-term yields will depend not least on whether the low-inflation environment proves robust towards future shocks to the economies.

The US 10-year yield is low in terms of the economic situation in the USA . High economic activity has meant that since June 2004 the Federal Reserve has raised the Fed Funds target rate from 1.0 per cent to 4.0 per cent, the level in November 2005. The long-term yield normally rises in periods when the Fed tightens monetary policy, cf. Chart 1, but on this occasion the long-term yield has remained virtually unchanged, so that the yield curve has become almost flat. This is remarkable as the Fed still characterises monetary policy as expansionary.

US SHORT-TERM AND LONG-TERM YIELDS	Chart 1
Note: Shaded areas indicate periods of significant tightening of monetary policy by the Federal Reserve. Source: EcoWin and own calculations.

The higher Fed Funds target rate has not impacted the long-term yield because the tightening had already been discounted in the market in June 2004, cf. Chart 2. Moreover, there has been slight downward adjustment of the estimates of US GDP growth in 2005 and 2006.

EXPECTATIONS OF THE FED	Chart 2
Note: The futures curve shows the market's expectation of the 3-month yield in December 2006 and is derived from the 3-month deposit future expiring in December 2006. Source: EcoWin and own calculations.

The interest-rate level is also low in the sense that for a prolonged period the long-term yield has been substantially below the market's long-term expectations of nominal growth in the USA , cf. Chart 3 (left-hand side). The long-term growth expectations normally indicate a minimum requirement of the yield on an investment in bonds with equivalent maturity. The long-term yield is now back at the level of the 1960s, before inflation began to accelerate, cf. Box 1 .

EXPECTED GROWTH AND LONG-TERM YIELD, AND FAIR-VALUE MODEL	Chart 3
Note: Left: The expected nominal GDP growth is calculated as the sum of the expected average annual real growth and inflation 10 years ahead stated as half-yearly observations. Right: Observations as of month-end. The model is the 10-year yield explained by core inflation (lagged 1 period), the official interest rate and the ISM index (lagged 1 period). Source: EcoWin, Consensus Economics and own calculations.

Fair-value models explain the long-term yield by a few fundamental macroeconomic variables. One example of this type of model is shown in Chart 3 (right-hand side) where the 10-year yield is explained by core inflation, the Fed Funds target rate and the ISM index (growth indicator). The actual and model-estimated yields are close to parallel up to 2004, after which the model indicates a rising yield due to the Fed's tightening of monetary policy. Today the long-term yield is around 100 basis points below the level indicated by the model's variables. Alternative simple model specifications lead to the same conclusion.

INFLATION RISK PREMIUM

INFLATION VOLATILITY	Box 2
Inflation volatility is calculated as Where p is actual annual inflation, E(p) is expected inflation, and n is the number of observations, in this case on a monthly basis (with no adjustment for overlapping observations). For the 10-year yield a specification is required of expected inflation in a 10-year perspective in order to measure the relevant inflation volatility. Each year in April and October Consensus Economics requests expected inflation on a 6-10-year horizon. These observations are available as from 1990. In addition, in the Survey of Professional Forecasters (SPF) the Fed and the ECB on a quarterly basis compile long-term inflation expectations for respectively the USA and the euro area. For the USA , these are available back to 1992, while they are available for the euro area as from 2001. Chart 6 (left-hand side) presents expectations of long-term and actual inflation. The fact that long-term inflation expectations are only available for the last few years is naturally an impediment. The unmistakeable sluggishness of inflation expectations inspires a simple model for long-term inflation expectations further back in time. The method is to use the HP filter to "smooth" actual inflation sufficiently to achieve a measure that captures the sluggishness of the observations of actual expected inflation in the period for which the observations are available. The principle is shown in Chart 6 (right-hand side) and corresponds to investors at a given point in time forming their long-term inflation expectations as a weighted average of actual previous and future inflation rates. This captures an element of foresight in the formation of expectations and in the present case the method appears to give a reasonable estimate in relation to the development in interest rates.  Inflation volatility is now calculated as a moving 60-month standard deviation for actual inflation in relation to the model's estimate of expected inflation. The actual length of the window in which this volatility is calculated does not change the conclusions. Inflation volatility for Germany is calculated in the same way.
ACTUAL AND EXPECTED INFLATION IN THE USA	Chart 6
Source: Consensus Economics, EcoWin and own calculations.

INFLATION VOLATILITY AND INTEREST RATES IN THE USA AND GERMANY	Chart 5
Source: EcoWin and own calculations.

The result indicates that the inflation risk premium is currently at a lower level than in e.g. the 1970s. If the low inflation and inflation volatility are expected to be sustained, then uncertainty concerning the real return on a nominal asset has declined, so that investors do not have to be compensated for the uncertainty of future inflation to the same degree as before. This is confirmed by the fact that the drop in interest rates in both the USA and Germany has been greater than can be explained by lower inflation in itself, as illustrated by the fact that the real interest rate – calculated as the 10-year yield less the actual rate of inflation – is also currently at a very low level in both countries.

From equation (1) irp = i – r – E(p). A rough measure of the size of the inflation risk premium can therefore be obtained by deducting the expected long-term real rate of return (approximated as expected real long-term GDP growth), and expected long-term inflation, from the nominal rate of interest.^[2] This measure indicates that since 1992 the inflation risk premium has fallen by approximately 2 per cent in Germany , and by slightly less in the USA .

A detailed study of the inflation risk premium in the US is found in Buraschi and Jiltsov (2005),^[3] where the inflation risk premium is estimated for the entire yield curve out to the 10-year maturity. The inflation risk premium for the 10-year yield rather closely mirrors the estimated inflation volatility for the USA in Chart 5 (left-hand side), but in terms of size it is slightly below the result of the above less detailed calculation. The inflation risk premium thus falls from a level around 120-140 basis points at the end of the 1970s and start of the 1980s to approximately 40 basis points at the end of the 1990s. The difference in the inflation risk premiums for short-term and long-term bonds, respectively, narrows considerably and is virtually eliminated at the end of the period, indicating the expectation of a future stable inflation environment.

As a deeper underlying cause of the low interest-rate level is thus not only low, but also stable inflation, the role of central banks can be highlighted. In recent years the western central banks have achieved the credibility to anchor inflation expectations at a low and stable level. In the future the central banks must be expected to continue to seek to maintain a stable inflation environment. Whether they succeed depends on whether the central banks' current monetary-policy strategies, combined with the overall economic policy, prove to be robust towards future shocks to the economies. 

GLOBALISATION

International trade and capital flows have expanded in recent years. This globalisation has served to push down price development and interest-rate levels in the western economies – directly via imports of cheaper goods (for both consumption and production) and indirectly by influencing wage formation.

The direct impact of globalisation should be viewed in the light of recent years' significant increase in the Asian countries' share of global trade, especially China . China 's trade with the USA has quadrupled over the last 10 years. Since 2003 prices for goods imported from Asia to the USA have generally fallen, while goods imported from the EU have become relatively more expensive, cf. Chart 7 (left-hand side), primarily due to the euro's strengthening against the dollar. Clothing and electronics in particular are imported from Asia, and the price development within these goods categories has been considerably more dampened than for consumer prices generally. However, these good categories carry only little weight in the overall index of consumer prices, so that the direct effect is limited, cf. Chart 7 (right-hand side). But other goods are imported from Asia , and the real effect on the development in consumer prices is probably greater.

US IMPORT AND CONSUMER PRICES	Chart 7
Note: Left: Newly-industrialised Asia includes Hong Kong , Singapore , South Korea and Taiwan . Southeast Asia includes Brunei , Cambodia , Indonesia , Laos , Malaysia , Myanmar , the Philippines , Singapore , Thailand and Vietnam . Right: CPI excludes energy and food. Weights from 1 January 2005 are used.
Source: Bureau of Labor Statistics, Bloomberg and own calculations.

It should be noted that compared to its output the USA 's international trade is limited. More open economies have seen a greater inflation impact on producer and consumer prices from lower import prices. Experience from e.g. Norway , which has lifted most of its import restrictions on textiles, shows a higher direct globalisation effect on price developments.

The indirect effect of globalisation is the increased wage competition between industrialised and developing countries. For example, the average hourly wage of a worker in China is approximately 15 per cent of the hourly wage of a US worker, and this gap is even more pronounced for other southeast Asian countries.[4] This has led to greater cost awareness in labour-intensive, highly competitive sectors in the west, as the competition makes it harder to raise prices in order to compensate for higher costs. This in itself ensures moderate price development.

Globalisation does not stop here. The downward pressure on prices and wages will continue, and contribute to lower inflation expectations. Especially via the indirect effect of greater wage competition from Asia , globalisation has already significantly dampened price developments, thereby contributing to the low level of interest rates in the western economies. The future impact of globalisation will depend on growth in the developing countries and the speed at which these countries approach western wage and price levels.

LOW INVESTMENTS

International economic conditions have been favourable in recent years, but this has not had the expected beneficial impact on investments. Viewed against global GDP growth the global investment ratio is thus relatively low, cf. Chart 8.^[5] The generally weak demand for investment loans in recent years is assessed to be a factor contributing to the low long-term (real) yields.

GLOBAL INVESTMENT RATIO AND ECONOMIC GROWTH	Chart 8
Source: IMF, World Economic Outlook database.

Realised global savings and investments are by definition identical.^[6] This means that actual global savings are also at a relatively low level. On the other hand, savings and investment plans can deviate considerably from each other, but changes in real interest rates and production will help to ensure equilibrium. If the propensity to save e.g. exceeds the propensity to invest, a falling real interest rate will help to balance savings and investments.

Recent decades' general tendency for global savings and investments to recede reflects the development in the industrialised countries, cf. Chart 9 (left-hand side). In the same period savings have tended to rise in the emerging markets and oil-producing economies, cf. Chart 9 (right-hand side), while investments in these countries rose up to the Asian crisis in 1997, after which they fell to a lower level.

SAVINGS AND INVESTMENTS	Chart 9
Source: IMF (2005).

The IMF (2005) concludes that the development in global savings in the period 1997-2004 can be explained by the factors that normally drive savings, e.g. lending to the private sector, where generally rising housing prices have led to an increase in the mortgage basis, and public-sector savings.

On the other hand, according to the IMF, the development in global investments in the period 1997-2004 cannot be explained by the factors that are normally assumed to determine investments. They ought to have risen both in industrialised countries and emerging markets, primarily as a result of decreasing financing costs.^[7] In Japan and the euro area, however, investments have continued to fall, as in previous decades, while in the USA investments have fluctuated around a long-term average of approximately 20 per cent of GDP. In emerging markets and oil-producing economies investments have fallen since the Asian crisis, even though investments in China have risen strongly and, like savings, constitute around half of its national GDP.

The global tendency for lower investment has limited the demand for lending and thereby contributed to reducing real interest rates. This development has been particularly apparent in Asia and in oil-producing economies, whose current accounts have improved significantly in recent years, cf. Chart 9 (right-hand side). Low investment is hardly a permanent situation, and a future increased investment level will probably exert upward pressure on the real interest rate. However, the development may also reflect a structural adjustment in terms of a reduction of the physical capital on the transition to more knowledge-intensive production.

PURCHASE OF GOVERNMENT BONDS BY PENSION COMPANIES AND ASIAN CENTRAL BANKS

Pension companies and the Asian central banks, which in current years demand large volumes of bonds, are two important groups of players in the bond markets. In the case of both groups, demand is derived from other factors besides real rates of return. With regard to pension companies, the demand is especially due to new rules for compiling pension companies' solvency and the covariation of long-term yields with the present value of the pension companies' obligations. The demand from Asian central banks has its roots in the exchange-rate policies of the Asian countries.

Pension companies
Gradual transition from defined benefit to defined contribution pension systems in the western countries has increased the importance of the pension companies' investment policies. At the same time, new accounting rules and changed supervisory practices have increased their demand for long-term bonds and thereby contributed to falling long-term yields. The implications of these new requirements are that interest-rate changes now affect the book values of both assets and liabilities. Under the new rules, pension companies wishing to minimise the risk exposure (volatility) of their financial results, and thereby the risk of insolvency, have a greater incentive to invest in assets showing a high degree of correlation with the valuation of future pension obligations – i.e. in bonds or swaps with long duration.

The valuation of future pension obligations is based on the discount rate. So far this has been assumed to be constant in some countries, e.g. in Europe, while in others, such as the USA , it has been fixed as a weighted portfolio average of historical returns on bonds and shares. The discount rate has thus in neither case reflected the current market conditions.

New accounting standards and regulations for European pension companies are now introducing market-based valuation of the obligations, so that the discount rate reflects movements in the yield curve. This requires that the discount rate is determined on the basis of the yield on a corporate bond of high credit quality (AA or the equivalent).

Previously, an interest-rate change did not alter the book value of the liabilities, while the assets appreciated if interest rates fell and depreciated if interest rates rose. Therefore solvency improved when interest rates fell. Under the new rules, the liabilities appreciate when interest rates are falling. In many cases the liabilities increase by more than the assets appreciate.^[8] The reason is that many companies have operated with a shorter duration for their assets than their liabilities, since historically liabilities have not been subject to value adjustment. As a consequence, many pension companies have increased their demand for duration, typically in the form of long-term bonds. 

In global terms pension companies are major players in the financial markets and therefore a large-scale restructuring of their portfolios can have decisive influence on price formation. In several countries pension companies' assets as a ratio of GDP are very substantial. A weighted average is 60 per cent for the OECD area (OECD 2005). The USA , the UK and the Netherlands , among others, are above the average, while most OECD countries are somewhat lower.

The effect of portfolio restructuring from shares and short-term bonds to long-term bonds as declining yields at the long-term end of the yield curve is already apparent. Globally, this can be expected to be further amplified in step with the full international implementation of the rules and expansion of the pension sector.

Asian central banks
Asian central banks' large-scale purchases of US government bonds are a popular explanation for the current low level of interest rates.

The US economy is currently characterised by substantial saving deficits. The deficit on the balance of payments thus accounted for 85 per cent of the global balance-of-payments deficit in 2004.^[9] Concurrently, in recent years a considerable deficit on US government finances has emerged, and thereby a higher bond-issuing requirement. Normally, a larger supply of bonds will, all other things being equal, lead to falling bond prices and thus to rising interest rates, but so far Asian central banks, especially those of Japan and China, are willing to take up vast amounts of government bonds, thereby contributing to lower interest rates. In 2004 the global accumulation of reserves exceeded 530 billion dollars. This is equivalent to approximately 80 per cent of the global balance-of-payments deficit being financed by central banks, predominantly from Asia .

The background to the bond purchases is especially the Asian countries' current-account surpluses, coupled with a preference for a stable or completely fixed exchange rate against the dollar.^[10] The fixed-exchange-rate policy is achieved primarily by buying dollars against the local currency, and to a great extent the increasing foreign-exchange reserves are placed in US bonds. The purchases are thus a derived effect of the economic policy, rather than an investment in order to achieve higher yields. This has nourished speculation that the central banks are willing to finance the global balance-of-payments deficit on other terms than e.g. private investors. 

It is very hard to quantify such effects, and in the economic literature opinion is divided on the impact of the bond purchases on interest rates, as well as the sustainability of a situation where the US deficits are increasingly financed by especially the People's Bank of China. In terms of impact on interest rates there is, however, agreement on the sign-effect, with the estimates varying between approximately 40 and 200 basis points lower interest rates as a consequence of the bond purchases. Studies of the interest-rate impact are impeded by the fact that accumulation of reserves is a relatively new phenomenon. 

For as long as they are financed, the large global imbalances can continue. If the Asian central banks finance the deficit on terms that other investors are less inclined to accept, then, potentially, interest rates may increase if the central banks at some point reduce or suddenly stop their bond purchases. The effect is uncertain, however, and will be less significant if there is a gradual transition to more flexible Asian exchange rates, possibly combined with a reduced borrowing requirement in the USA, e.g. via a fiscal tightening. A possible transition to more flexible Asian exchange rates, which will probably result in the revaluation of local Asian currencies, is complicated particularly by the considerable capital losses that the Asian countries would in such case incur on their very large portfolios of dollar-denominated assets. 

PERSPECTIVES

The current low level of interest rates should be viewed in the light of low inflation expectations, a low inflation risk premium, increased globalisation, low global investments and large-scale bond purchases by pension companies and Asian central banks. In isolated terms, continued focus on fighting inflation and increasing globalisation can contribute to keeping long-term yields lower than experience from the preceding 30-40 years would indicate for some time to come. Bond purchases by pension companies and Asian central banks, as well as low investments, are assessed to be temporary phenomena that will disappear in the longer term. When that happens, it will tend to push interest rates up.

LITERATURE

Buraschi, Andrea and Alexei Jiltsov (2005), Inflation risk premia and the expectations hypothesis, Journal of Financial Economics 75, pp. 429-490.

Cappiello, Lorenzo and Stéphane Guéné (2005), Measuring market and inflation risk premia in France and in Germany , ECB Working Paper no. 436.

The Danish Economic Council (2004), The Danish Economy, International Outsourcing, Chapter II, autumn (in Danish with English summary).

Evans, M.D. (2003), Real risk, inflation risk and the term structure. The Economic Journal 113, pp. 345-389.

Genberg, Hans, Yung Chul Park and Avinash Persaud (2005), Official Reserves and Currency Management in Asia : Myth, Reality and the Future, CEPR Conference draft of the 7th Geneva Report on the World Economy.

Higgins, M. and T. Klitgaard (2004), Reserve Accumulation: Implications for Global Capital Flows and Financial Markets, Current Issues in Economics and Finance, Vol. 10, No. 10, Federal Reserve Bank of New York.

IMF (2004), Global Financial Stability Report, Risk Management and the Pension Fund Industry.

IMF (2005), World Economic Outlook, Chapter 2, September.

OECD (2005), Pension Markets in Focus, June, issue 1.

Rogoff, K. (2003), Globalization and Global Disinflation, IMF, Paper prepared for the Federal Reserve Bank of Kansas City conference on "Monetary Policy and Uncertainty: Adapting to a Changing Economy" Jackson Hole, WY.

Roubini, N. and B. Setser (2005), Will the Bretton Woods 2 Regime Unravel Soon? The risk of Hard Landing in 2005-2006, Article for the symposium "Revived Bretton Woods System: A New Paradigm for Asian Development?", Federal Reserve Bank of San Francisco and UC Berkeley, San Francisco, February 2005.

Wagner, H. (2001), Implications of Globalization for Monetary Policy, IMF Working Paper 01/184.

---

[1]  In general terms, risk premium = price of risk × risk factor. Inflation volatility corresponds to considering the risk factor. The price of risk will generally vary over time and depend positively on the market's risk aversion. It is reasonable to assume that the risk factor determines the risk premium in the longer term.

[2]  These expectations are compiled by Consensus Economics, and the sum is shown in Chart 3 (left-hand side) for the USA . 

[3]  So far empirical studies of inflation risk premiums are rather limited. Other studies include Evans (2003) and Cappiello and Guéné (2005).

[4]  The wage differences cannot be explained solely by productivity differences, since to some extent they are adjusted for educational background, experience, etc. (The Danish Economic Council, 2004).

[5]  The concepts "investment" and "savings" in this section are to be understood as ratios of GDP. The concepts used are gross measures in current prices. 

[6] This expresses that in principle the sum of the national current accounts of the balance of payments is zero. Due to measurement uncertainty this is not the case in practice, however. The difference between global savings and investment in this section is moreover a consequence of the fact that the IMF figures used only almost cover the entire world.

[7] One explanation for the model's over-prediction may be that it does not take sufficient account of falling prices for IT investments (investments are measured in current prices). In volume terms the drop in investments is thus more moderate (IMF, 2005).

[8]  A realistic duration gap between assets and liabilities could be 10 years, e.g. the case of the Netherlands , where the average pension company has a bond portfolio with a duration of 5 years, while the duration of the liabilities is 15 years.

[9]  The figures in this section refer to Roubini and Setser (2005).

[10] There has been special focus on China 's exchange-rate policy, which was previously a fixed-exchange-rate policy against the dollar exclusively, but in summer 2005 was changed to a fixed-exchange-rate policy via-à-vis a basket of currencies in which the dollar is included with a substantial weight.