Prasad V. Bidarkota

Commodity Prices and the Terms of Trade

On combining national terms-of-trade data for developing countries with world prices of internationally traded primary commodities, it is found that variation in the world prices of three or fewer key exported commodities account for 50% or more of the annual variation in the terms of trade of a typical developing country. A considerable fraction of the variation is specific to a particular commodity and, given that the overall importance of primary commodities differs across developing countries, it is possible to account for much of the heterogeneity across them. It is concluded that commodity price fluctuations should be central features of two related literatures: studies of business cycle transmission across developing and industrialized nations, and empirical work aimed at constructing perpetual claims on developing country incomes as suggested by Shiller in 1995.

Optimal univariate inflation forecasting with symmetric stable shocks

Monthly inflation in the United States indicates non-normality in the form of either occasional big shocks or marked changes in the level of the series. We develop a univariate state space model with symmetric stable shocks for this series. The non-Gaussian model is estimated by the Sorenson-Alspach filtering algorithm. Even after removing conditional heteroscedasticity, normality is rejected in favour of a stable distribution with exponent 1·83. Our model can be used for forecasting future inflation, and to simulate historical inflation forecasts conditional on the history of inflation. Relative to the Gaussian model, the stable model accounts for outliers and level shifts better, provides tighter estimates of trend inflation, and gives more realistic assessment of uncertainty during confusing episodes.

Asymmetries in the Conditional Mean Dynamics of Real GNP: Robust Evidence

We investigate asymmetries in the conditional mean dynamics of U.S. GNP. Because the statistical evidence on nonlinearities in the conditional mean could be influenced by the presence of outliers or by a failure to model conditional heter oske dasticity, we explicitly account for outliers by assuming that the innovations are drawn from the stable family, and model time-varying volatility by a GARCH(1, 1) process. We also allow for the possibility of long memory in the series with fractional differencing. Our results indicate statistically significant nonlinearities in the conditional mean that persist even after accounting for these features in the data.

The comparative forecast performance of univariate and multivariate models: an application to real interest rate forecasting

Abstract Does the use of information on the past history of the nominal interest rates and inflation entail improvement in forecasts of the ex ante real interest rate over its forecasts obtained from using just the past history of the realized real interest rates? To answer this question we set up a univariate unobserved components model for the realized real interest rates and a bivariate model for the nominal rate and inflation which imposes cointegration restrictions between them. The two models are estimated under normality with the Kalman filter. It is found that the error-correction model provides more accurate one-period ahead forecasts of the real rate within the estimation sample whereas the unobserved components model yields forecasts with smaller forecast variances. In the post-sample period, the forecasts from the bivariate model are not only more accurate but also have tighter confidence bounds than the forecasts from the unobserved components model.

Testing for Persistence in Stock Returns with GARCH-Stable Shocks

We investigate persistence in CRSP monthly excess stock returns, using a state space model with stable disturbances. The non-Gaussian state space model with volatility persistence is estimated by maximum likelihood, using the optimal filtering algorithm given by Sorenson and Alspach (1971 Automatica 7 465-79). The conditional distribution has a stable α of 1.89, and normality is strongly rejected even after accounting for GARCH. However, stock returns do not contain a significant mean-reverting component. The optimal predictor is the unconditional expectation of the series, which we estimate to be 9.8% per annum.

Consumption asset pricing with stable shocks--exploring a solution and its implications for mean equity returns

Abstract We study the consumption based asset pricing model due to Lucas (Econometrica 46 (1978) 1429). The exogenous endowment sequence is modeled as a linear stochastic process driven by stable shocks in an otherwise standard framework. The Gaussian process emerges as a special case. We derive exact analytical solutions for asset prices and returns, and provide conditions under which these exist. We also study the ability of the model to generate realistic values of observed mean rates of return.

Sectoral Investigation of Asymmetries in the Conditional Mean Dynamics of the Real U.S. GDP

We investigate asymmetries in the conditional mean dynamics of four sectors of the U.S. GDP data. Since the statistical evidence on nonlinearities in the conditional mean could be influenced by the presence of outliers, or by a failure to model conditional heteroskedasticity, we explicitly account for outliers by assuming that the innovations are drawn from the stable family, and model time-varying volatility by a GARCH(1,1) process. We also allow for the possibility of long memory in the series with fractional differencing. Our results indicate only weak evidence of significant nonlinearities in the conditional mean in some sectors of the GDP.

Alternative Regime Switching Models for Forecasting Inflation

US inflation appears to undergo shifts in its mean level and variability. We evaluate the performance of three useful models for capturing such shifts. The models studied are the Markov switching models, state space models with heavy-tailed errors, and state space models with compound error distributions. Our study shows that all three models have very similar performance when evaluated in terms of the mean squared or mean absolute forecast errors. However, the latter two models are considerably more parsimonious, and easily beat the more profligately parameterized Markov switching models in terms of model selection criteria, such as the AIC or the SBC. Thus, these may serve as useful continuous alternatives to the popular discrete Markov switching models for capturing shifts in time series. Copyright

Forecast Performance of Neural Networks and Business Cycle Asymmetries

Forecast performance of artificial neural network models are investigated using Ashley et al. (1980) and the neural network nonlinearity test proposed by Lee et al. (1993) is employed to find possible existence of business cycle asymmetries in Canada, France, Japan, UK and USA real GDP growth rates. The results show that neural network models are more accurate than linear models for in-sample forecasts. However, when comparing the out-of-sample, linear models performed better than neural network models in all series. Results from neural network tests show that business cycle asymmetries do prevail in all the series.

Intrinsic Bubbles And Fat Tails In Stock Prices: A Note

We study the constant discount rate present value model for stock pricing in a stochastic setting where the exogenous dividend stream is modeled as a random walk with innovations drawn from the family of stable distributions. We derive an exact analytical solution for the fundamental stock price. We evaluate the ability of the model fundamentals and the dividends-driven intrinsic bubbles to explain the observed variation in annual U.S. stock prices. We compare results obtained in this setting with those from the traditional model where all stochastic processes are driven by Gaussian shocks.