United States of America: Selected Issues

This Selected Issues paper on the United States analyzes problems in the measurement of output and prices. The paper examines income versus expenditure measures of national output. Sources of consumer price index and findings of the Boskin Commission are discussed, and mismeasurement of output and productivity is analyzed. Developments in productivity across industries in the United States are described. In particular, the paper focuses on the slowdown in aggregate productivity growth that began in the mid-1970s and examines whether this slowdown has continued in recent years and is common across industries.

Abstract

This Selected Issues paper on the United States analyzes problems in the measurement of output and prices. The paper examines income versus expenditure measures of national output. Sources of consumer price index and findings of the Boskin Commission are discussed, and mismeasurement of output and productivity is analyzed. Developments in productivity across industries in the United States are described. In particular, the paper focuses on the slowdown in aggregate productivity growth that began in the mid-1970s and examines whether this slowdown has continued in recent years and is common across industries.

VI. The Recent Behavior of Stock Prices1

1. Stock market prices have risen sharply over the past two years, reaching all-time highs and raising concerns that “irrational exuberance has unduly escalated asset values.”2 With the rise in prices, price/earnings ratios have moved out of line with historic averages and dividend/price ratios (also referred to as dividend yields) have fallen to historic lows, suggesting that stock prices are high relative to economic fundamentals (Chart 1). To examine this question, a simple model explaining the dividend/price ratio in terms of fundamental factors was estimated, and the model’s ability to predict the movements in the dividend yield over the period 1995–96 was analyzed. One version of the model can account for most of the decline in the dividend yield over the last two years.

CHART 1
CHART 1

UNITED STATES: STOCK PRICES, DIVIDENDS, AND MUTUAL-FUND INFLOWS

Citation: IMF Staff Country Reports 1997, 097; 10.5089/9781451839494.002.A006

Sources: The Wall Street Journal and the Financial Times; Standard & Poor’s, a division of McGraw-Hill; and Investment Company Institute.

2. The dividend/price ratio reflects the market’s forecast of future dividend growth and risk premia.3 Thus, variables that either forecast future dividends or are related to risk premia should help explain changes in the dividend yield. In addition, substantial net purchases of mutual funds have taken place in the past few years. These mutual fond inflows may reflect a change in longer-term fundamentals that is not captured in the other economic variables. Hence, two models for the dividend yield are estimated: a basic version, and an augmented one that includes a measure for mutual fund inflows.

3. Research suggests that variables like the yield curve, real interest rates, inflation, and the default premium are related to future economic activity, discount rates, and risk premia.4 In the basic model, the log of the dividend yield (LDP) is related to the slope of the yield curve (YC) (which represents the difference between long-term and short-term interest rates); the default premium (DEF) (measured as the difference in yields on corporate and government bonds); the (ex-post) real rate of interest on three-month Treasury bills (RTB); and the CPI inflation rate (PI). In the augmented version of the model, net purchases of mutual funds (MUFT) from the Investment Company Institute is added as an explanatory variable.

4. For each version, a model was estimated with two lags of the log dividend yield and a contemporaneous value and two lags of each explanatory variable. Each model was fitted to the period March 1984 to December 1994.5 Forecasts were then derived from the two models for the period January 1995 to February 1997. The long-run version of each model is as follows (standard errors in parentheses):

Basic model:

log(D/P)=0.36+0.07DEF+0.10PI+0.09RTB+0.08YC(0.12)(0.06)(0.01)(0.01)(0.02)

Augmented model:

log(D/P)=0.77-0.04DEF+0.06PI+0.06RTB+0.07YC-0.21MUFI6(0.29)(0.10)(0.02)(0.02)(0.02)(0.11)

In the dynamic versions of both models, each variable except DBF was statistically significant at the 1 percent level (when testing for the significance of all lags). DEF was significant at the 6 percent level in the first model, and at the 11 percent level in the second.

5. Charts 2 and 3 show the forecast performance of each model over the January 1995-February 1997 period.7 As can be seen, the basic model does not perform well in explaining recent movements in the dividend yield. Although the basic model forecasts a decline in the yield, the actual decline was much larger over the forecast period (the root mean square error (RMSE) of the forecast was 0.22). In particular, after mid-1995, actual values for the dividend yield lie outside the two-standard error confidence interval for the forecast. In contrast, the augmented model closely tracks the movements in the dividend yield (the RMSE of the forecast was 0.07). The actual dividend yield is generally within the two-standard error band for the forecast (Chart 3). These results imply that mutual fund inflows are important in explaining the decline in the dividend yield in the recent period.8

CHART 2
CHART 2

UNITED STATES: LOG DIVIDEND YIELD: FORECAST AND ACTUAL (BASIC MODEL)

Citation: IMF Staff Country Reports 1997, 097; 10.5089/9781451839494.002.A006

Sources: Standard & Poor’s, a division of McGraw-Hill; and Fund staff estimates.
CHART 3
CHART 3

UNITED STATES: LOG DIVIDEND YIELD: FORECAST AND ACTUAL (AUGMENTED MODEL)

Citation: IMF Staff Country Reports 1997, 097; 10.5089/9781451839494.002.A006

Sources: Standard & Poor’s, a division of McGraw-Hill; and Fund staff estimates.

6. A forecast attribution was performed using the augmented model. The intent was to measure how much each variable contributed to the change in the forecast, and by implication to estimate how important each variable was in explaining the actual decline in the dividend yield.9 T can explain only about he table below shows the forecast attributions. Most of the variables other than mutual-fund inflows have fairly low explanatory power in the forecast. The basic model, which excludes mutual fund inflows, can explain only about 15 percent of the decline in the log dividend yield, while the augmented model explains about 80 percent of the decline. In the augmented model, the mutual fund inflows variable explains nearly 75 percent of the change in the dividend yield.

Forecast Attribution

(In percent)

article image

7. The inflow of money into mutual funds may reflect a shift in portfolio allocations by individual investors. The development of mutual funds in recent years has made it easier for individuals to hold a diversified portfolio of stocks and has significantly lowered transactions costs. Historically, stock holdings have been highly concentrated among a relatively small number of wealthy households. More recently, however, more households have been participating in the stock market, particularly through mutual funds.10 The previously highly concentrated holdings of stocks could have depressed stock prices relative to what they would have been if holdings were more uniformly distributed.11 If the increase in mutual-fund in flows represents a shift of portfolios toward a less-concentrated allocation, then the recent rise in prices and decline in yields may represent a move to a different equilibrium, rather than a departure from equilibrium. Alternatively, if expectations of future returns on the scale of those experienced in the last two years have motivated recent inflows to mutual funds, then the evidence in this paper might be consistent with “irrational exuberance.”

List of References

  • Basak, Suleyman, and Domenico Cuoco, 1997, “An Equilibrium Model with Restricted Stock Market Participation,Rodney L. White Center for Financial Research Paper # 1–97 (Philadelphia: Wharton, University of Pennsylvania).

    • Search Google Scholar
    • Export Citation
  • Frost, Robin, 1996, “A Look in the Mirror: The Changing Face of Stock Investors,Wall Street Journal Interactive Edition, May 28.

  • Kramer, Charles, 1996, “Stock-Market Equilibrium and the Dividend Yield,IMF Working Paper WP/96/90 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Laderman, Elizabeth, 1997, “Deposits and Demographics?”, Federal Reserve Bank of San Francisco Economic Letter, Number 97–19, June 27, 1997.

    • Search Google Scholar
    • Export Citation
1

Prepared by Charles Kramer.

2

Remarks by Federal Reserve Board Chairman Alan Greenspan at the American Enterprise Institute, December 5, 1996.

3

For a complete explanation of the theoretical model used in this analysis and its empirical specification and estimation, see Kramer (1996).

4

These variables are often used in empirical studies of asset pricing, and they forecast dividend growth rates and stock returns out of sample.

5

The data on mutual-fund inflows are not available before 1984.

6

The coefficient and standard error for MUFI are multiplied by 104 for ease of presentation.

7

The forecasts presented here are dynamic forecasts, while those in Kramer (1996) are onestep-ahead static forecasts.

8

The difference between the forecasts from the basic and augmented models is not accounted for by differences in the coefficients on the common variables in the two models. If the mutual fond inflow variable is held constant at its December 1994 value over the forecast horizon, the forecast from the augmented model is similar to the one from the basic model.

9

The contribution of each variable is calculated as the effect on the forecast of the change in the dividend yield that results from allowing the variable to vary over the forecast period, expressed as a percent of the actual change in the dividend yield. The contributions do not add to 100 percent because of the lags in the independent variables and because part of the change in the forecast is explained by historical residuals.

11

Such a phenomenon is consistent with the “equity premium puzzle” that historical stock returns have been too high to be consistent with standard equilibrium models (see Basak and Cuoco (1997)).

United States: Selected Issues
Author: International Monetary Fund
  • View in gallery

    UNITED STATES: STOCK PRICES, DIVIDENDS, AND MUTUAL-FUND INFLOWS

  • View in gallery

    UNITED STATES: LOG DIVIDEND YIELD: FORECAST AND ACTUAL (BASIC MODEL)

  • View in gallery

    UNITED STATES: LOG DIVIDEND YIELD: FORECAST AND ACTUAL (AUGMENTED MODEL)