Debunking the Fed Model

Includes: AGG, SPY
by: Charlie Fell

John Burr Williams wrote in his 1938 classic, The Theory of Investment Value, that “investment analysis measures the relative rather than the absolute value of any stock, and leaves to the economist the broad question of whether stocks are selling too high or too low.” More than seven decades later, and relative valuation continues to be the most popular technique employed by investment professionals.

One of the more dubious methods used to assess the relative attractiveness of the stock market is to compare the major market average’s dividend or earnings yield with the current level of long-term interest rates. The so called "Fed model" received almost universal approval as the Holy Grail of valuation following Alan Greenspan’s Humphrey-Hawkins testimony to Congress on July 22, 1997.

The report that accompanied the former Fed Chairman’s testimony revealed a strong positive correlation between the ten-year Treasury yield and the S&P 500’s earnings yield; the commentary noted that changes in the price/earnings multiple – the inverse of the earnings yield – “have often been inversely related to changes in long-term Treasury yields.”

The Federal Reserve’s presumed approval of the valuation tool emboldened the bullish strategists on Wall Street. The dividend yield had traditionally been used as the key variable employed for comparison with long-term interest rates, but by the mid-90s, this version of the model made stocks look anything but cheap. The solution was to replace the dividend yield with the earnings yield and to include forward-looking estimates instead of historic profit numbers. The use of opinion over fact however, enabled Wall Street’s sales machine to adjust their inputs at will, and ensure that the stock market never looked overpriced. Alan Greenspan's supposed validation of the model only served to increase its appeal and misuse.

The use of overly optimistic numbers may be damning enough, but the harshest criticism of the "Fed model" is that it is theoretically invalid and more to the point, simply does not work in practice. Indeed, the price that investors are willing to pay for a dollar of earnings is in secular decline, even though the yield on ten-year Treasuries has dropped to just 2 ½ per cent.

The earnings yield based on one-year forecast numbers is six percentage points above the 2 ½ per cent equilibrium value implied by the Fed model, and long-term interest rates are only barely higher than the forecast dividend yield. The widely-accepted positive correlation between the dividend yield and long-term interest rates has crumbled and turned negative. Indeed, the coefficient of correlation since the end of 1999 has been an impressive –0.80 with statistical significance.

What explains the spectacular breakdown in the Fed model? The most obvious answer is that it is theoretically flawed from the outset. Secular shifts in long-term Treasury yields are typically caused by sizable movements in inflation expectations, but inflation per se should have no impact on the valuation of stocks since they are a claim on real cash flows. However, the risk premium that investors require for equity investment increases, as the economy moves away from price stability towards either an inflationary or deflationary regime.

The verdict of history suggests that investors are willing to pay the most for a dollar of earnings when the inflation rate is contained between two and four per cent. Outside of this range and valuation multiples fall and yield measures rise. Thus, stock yields based on either dividends or earnings tend to be positively correlated with bond yields when inflation expectations move towards or away from an inflationary regime, and exhibit a negative correlation when inflation expectations move towards or away from a deflationary regime.

Why do inflation expectations matter? Prices provide consumers and producers with the information they need to assess the relative value of goods and services, which ensures that resources are allocated to their best uses. However, the information content in prices declines significantly when the aggregate price level changes unpredictably and during deflationary and inflationary regimes it becomes difficult to distinguish between relative prices that are advantageous and those that are disadvantageous. The historical evidence shows that both deflationary and inflationary regimes lead to greater economic volatility and consequently, individuals require additional compensation for investment in stocks or higher dividend yields and lower price/earning multiples.

The stock market’s dividend yield was routinely above long-term interest rates throughout the 19th century and through the first half of the 20th century, so the current narrow gap can hardly be described as an outlier. The apparent anomaly persisted due to high levels of economic volatility, and a business climate that apart from occasional war-induced inflations, was notable for its periodic bouts of deflation.

The established relationship persisted until the autumn of 1958 when the dividend yield on stocks fell below the yields available on long-term Treasury bonds for the first time since the heady days of 1929, a crossover that lasted only briefly. The crossover occurred as investors appreciated that the preceding recession was the first time post-World War II that an economic downturn had not been accompanied or followed by a declining price level. The “Great Depression” was not set to return; inflation was now the norm and stocks were considered to be the most effective inflation hedge.

Stocks prospered through the mid-1960s as the inflation rate was contained within the two to four per cent range and economic volatility subsided. The market began to struggle in the late-1960s and early-1970s however, as accommodative monetary policy, expansive fiscal policy and a collapse in the Bretton Woods system of fixed exchange rates all contributed to runaway inflation. The positive correlation between stock and bond returns that has come to denominate investment thinking became conventional wisdom.

The yield on Treasury bonds jumped from five per cent in the mid-1960s to 14 per cent in 1982, while the dividend yield on stocks increased from below three to almost seven per cent. The risk premium attached to stocks rose due to greater economic volatility, while upside inflation surprises saw bonds register negative real returns of more than three per cent per annum over the period.

The positive correlation continued during the 1980s and 1990s as the “Great Inflation” gave way to disinflation and a return to price stability. Both stocks and bonds enjoyed impressive bull markets, but they parted company more than a decade ago, and the correlation turned negative as deflation concerns came to the fore. The US economy is just one recession away from deflation and consequently, the flawed "Fed model" will continue to fail.

Disclosure: No positions