This June 1995 report from Salomon Brothers addressed a fundamental confusion in fixed income markets: the distinction between what forward rates imply about future spot rates and what the market actually expects. The document is 23 pages and presents algebraic relationships, theoretical frameworks, and empirical evidence that challenges the pure expectations hypothesis.
The report establishes that forward rates serve as break-even rates. When the spot curve is upward sloping, forward rates imply rising future spot rates. This represents the rate path that would make all maturities earn identical holding-period returns. The author emphasizes that this is a mathematical relationship, not a market forecast. An upward-sloping curve could reflect positive bond risk premia rather than expectations of rising rates.
The pure expectations hypothesis states that all government bonds have equal near-term expected returns, meaning all bond risk premia equal zero. Under this framework, forward-implied rate changes would exactly match expected rate changes. The report presents evidence that this hypothesis fails empirically. Forward rates prove to be poor predictors of actual rate changes, and long-term rates tend to move opposite to the direction implied by forwards.
The document contrasts forward-implied expectations with survey-based expectations. Surveys of bond market participants conducted by the Wall Street Journal reveal systematic differences. Forward rates consistently suggest larger expected rate increases than do surveys, indicating that forward rates are upward-biased measures of market expectations. This bias reflects positive bond risk premia demanded by investors for bearing duration risk.
The empirical section analyzes 1970-1994 monthly Treasury data on forward rates' forecasting ability. The report finds that steep yield curves predict high subsequent bond returns but not rising spot rates. When the curve is steep, forward rates imply substantial rate increases, yet realized rate changes average near zero. The implication is that the curve steepness reflects risk compensation rather than rate expectations.
On curve flattening positions, the report shows how to compute break-even yield changes for barbell versus bullet strategies. Figure 3 demonstrates that when the spot curve is concave, forwards imply future curve flattening. This occurs because intermediate-maturity bonds have higher yields relative to duration-matched barbells, requiring a flattening to equalize returns. The break-even spread change depends on both the initial yield differential and the rolldown tendencies of the barbell and bullet components.
The report was published in June 1995, one month after Part 1. The timing remained influenced by the 1994 bond market selloff. The Federal Reserve had completed its rate hiking cycle, and participants were evaluating whether the forward curve's implied rate path would materialize. The report's emphasis on distinguishing implied changes from expected changes addressed confusion that likely intensified during the volatile period.
The survey data referenced in the report comes from the Wall Street Journal's semiannual survey of Wall Street economists and analysts conducted 27 times between December 1981 and December 1994. In each survey, participants predicted the three-month Treasury bill rate and 30-year bond yield at the end of the next June or December. The report uses survey medians to represent market expectations.
The report documents that forward rates fail as rate predictors. Figure 7 evaluates forwards' ability to predict monthly rate changes from 1970-1994. The correlation between forward-implied rate changes and realized rate changes is positive but weak. More striking is the bias: forwards systematically overpredict rate increases. The average forecast error is negative, indicating that realized rate changes fall short of what forwards implied.
The comparison with survey expectations reveals the nature of this bias. Figure 8 shows forward-implied and survey-expected three-month Treasury bill rates six months ahead. The forward rates consistently lie above survey expectations. Figure 9 presents similar results for yield curve steepness expectations. The forwards imply more curve flattening than surveys suggest, again reflecting positive term premia rather than actual flattening expectations.
The report presents evidence that steep curves predict high bond returns. When the curve is steep, intermediate and long-term bonds outperform short-term bonds over the subsequent year. This occurs because the curve fails to flatten as much as forwards implied. The high yields on longer-duration bonds reflect risk compensation that gets realized as return rather than offset by capital losses.
The concave shape of the spot curve plays a role in understanding risk premia. The report notes that the typical spot curve shape is upward sloping and concave. This shape implies that bullet positions have higher yields than duration-matched barbells, and that forwards imply future curve flattening. The persistence of this shape over time suggests that investors demand higher yields for intermediate-maturity exposure, possibly related to supply-demand imbalances or hedging demands.
The report's framework for understanding the gap between forward-implied and expected rate changes informs current ETF-based strategies. The evidence that steep curves predict high returns rather than rising rates supports systematic approaches that increase duration exposure when the curve steepens.
Curve Steepness Signal for Duration Extension
A rule-based strategy could monitor the 2s10s Treasury slope and increase allocation to intermediate duration ETFs when the curve steepens beyond historical percentiles. The report's evidence suggests that steep curves reflect elevated term premia rather than rate expectations, making extension profitable on average.
Implementation via common ETFs: Monitor the spread between 10-year and 2-year Treasury yields. When the spread exceeds the 75th percentile of trailing 12-month observations, increase exposure from short-duration ETFs like SHY (1-3 year Treasuries, duration ~2) to intermediate ETFs like IEF (7-10 year Treasuries, duration ~7). The position would harvest the term premium implied by the steep curve.
Limitations include the need to establish thresholds for curve steepness signals and determine position sizing. The report provides conceptual support but no backtested implementation details. Transaction costs and the risk of further curve steepening during the holding period must be considered.
Survey-Adjusted Forward Rate Trades
The report's comparison of forward rates and survey expectations suggests an alternative approach: when the gap between forward-implied rate changes and survey expectations widens, position for less curve flattening than forwards imply.
Implementation via common ETFs: Calculate the forward-implied three-month rate six months ahead using current Treasury yields. Compare to available survey forecasts (Federal Reserve Bank of Philadelphia Survey of Professional Forecasters provides quarterly expectations). When forwards imply rate increases 50+ basis points larger than surveys suggest, construct a barbell position that benefits if the curve flattens less than implied.
A barbell combines short-duration (SHY) and long-duration (TLT, 20+ year Treasuries, duration ~17) to match the duration of an intermediate bullet (IEF). The barbell has lower yield but higher convexity. If the curve fails to flatten as much as forwards imply, the barbell underperforms but loses less than implied by the initial yield disadvantage.
Limitations include the irregular timing of survey releases and the difficulty of sizing positions based on the magnitude of the forward-survey gap. The report demonstrates the phenomenon but provides no quantitative rule for trade sizing.
Dynamic Roll-Down Harvesting
The report's analysis of rolling yields motivates strategies that systematically harvest roll-down return. The report's Figure 11 illustrates that when the curve is upward sloping, bonds earn their yield plus a roll-down return as they age down the curve.
Implementation via common ETFs: Maintain exposure to a constant maturity sector of the curve that exhibits consistent positive slope. ETFs like VGIT (3-10 year Treasuries, duration ~6) rebalance monthly to maintain target duration, automatically harvesting roll-down as constituent bonds age.
The roll-down return equals the duration-at-horizon multiplied by the slope of the curve segment. For a 7-year Treasury rolling down to 6 years over 12 months, with duration of 6 and curve slope of 30 basis points, the roll-down return approximates 1.8%. This adds to the bond's yield to produce the rolling yield shown in the report's forward rate curves.
Limitations center on the fact that ETF rebalancing differs from holding individual bonds. The ETF maintains constant duration by selling aged bonds and buying new ones, which creates tracking differences from a pure roll-down strategy. The tax efficiency is better for long-term holders since ETF rebalancing occurs within the fund.
The empirical analysis spans 1968-1995, a period that includes multiple interest rate regimes. The report does not test whether the relationships between curve shape and subsequent returns vary across regimes. The 1970s inflation era, the Volcker disinflation, and the declining rate environment of the early 1990s may have different risk premium dynamics.
The survey data comes from semiannual Wall Street Journal polls. The report does not address the representativeness of survey respondents or whether survey responses reflect genuine expectations versus stated views. Market participants may provide conservative public forecasts while acting on different private beliefs.
The analysis focuses on Treasury securities. The report does not extend the framework to corporate bonds, where credit risk creates additional return drivers beyond interest rate risk. The relationship between forward credit spreads and expected spread changes may differ from that of Treasury forward rates and expected rate changes.
The practical implementation details are limited. The report describes curve-flattening positions using barbells versus bullets but provides no guidance on how to select specific maturities, position sizing based on signal strength, or risk management protocols. The framework is conceptual rather than operational.
The distinction between forward-implied rate changes and market expectations remains central to fixed income analysis. Every modern term structure model must specify how to decompose forward rates into expected rate changes and risk premia. The report's framework appears in academic research and practitioner guides three decades later.
The empirical finding that forward rates are upward-biased predictors of rate changes continues to hold. Research published in the 2000s and 2010s using longer sample periods confirms that steep curves predict high bond returns but not rising rates. The risk premium interpretation has become standard.
The comparison of forward rates to survey expectations provides ongoing value. The Federal Reserve Bank of Philadelphia's Survey of Professional Forecasters continues to collect interest rate expectations quarterly. Market participants still compare forward curves to survey forecasts to gauge the magnitude of implied risk premia.
The concept of break-even rate changes for curve trades remains in active use. Portfolio managers constructing barbell-bullet positions calculate the curve flattening required to make the trades break even, just as the report demonstrates. The algebra connecting forward rates, rolling yields, and break-even yield changes continues to be taught.
For retail investors using ETFs, the report's insights about the relationship between curve shape and expected returns remain applicable. The framework explains why duration extension during steep curve environments has historically been profitable. The gap between forward-implied flattening and actual flattening creates opportunities to harvest term premia.
The report established that forward rates contain information about risk premia rather than pure expectations. This insight drives modern factor investing in fixed income. Strategies based on curve signals, carry measures, and term premium estimates all build on the foundation that forward rates reflect compensation for risk rather than forecasts of future spot rates.
The 1995 report succeeded in clarifying a source of confusion that persists in fixed income markets. Participants continue to conflate forward-implied rate changes with expected rate changes, leading to misinterpretation of curve signals. The report's patient explanation of the distinction between break-even rates and forecasts remains valuable for practitioners decades later.