Forecasting the Australian economy with DSGE and BVAR models

Reflecting the importance of commodities for the Australian economy, we construct a dynamic stochastic general equilibrium (DSGE) model of the Australian economy with a commodity sector. We assess whether its forecasts can be improved by using it as a prior for an empirical Bayesian vector autoregression (BVAR). We find that the forecasts from the BVAR tend to be more accurate than those from the DSGE model. Nevertheless, for output growth these forecasts do not outperform benchmark models, such as a small open economy BVAR estimated using the standard priors for forecasting. A Bayesian factor augmented vector autoregression produces the most accurate near-term inflation forecasts.     

Examining impulse response functions in cointegrated systems

A system of reduced forms with cointegrated variables may be estimated in two ways: as a vector autoregression in levels, or as a vector error correction model. The latter is a restricted version of the former. If there is cointegration, imposing this restriction will yield more efficient estimates. However, at short horizons, vector error correction estimates are known to perform poorly relative to those from a vector autoregression. We examine how this property affects impulse response functions. A Monte Carlo experiment, and an example, suggest that impulse response functions of the two models are similar at short horizons, but different at long horizons. This suggests that the loss of efficiency from vector autoregression estimation is not critical at the commonly used short horizon. Our results complement parallel arguments focusing on forecast errors made by Clements and Hendry (1995), Hoffman and Rasche (1996), and Lin and Tsay (1996).     

HOW IMPORTANT ARE TERMS‐OF‐TRADE SHOCKS?

According to conventional wisdom, terms‐of‐trade shocks represent a major source of business cycles in emerging and poor countries. This view is largely based on the analysis of calibrated business‐cycle models. We argue that the view that emerges from empirical structural vector autoregression (SVAR) models is strikingly different. We estimate country‐specific SVARs using data from 38 countries and find that terms‐of‐trade shocks explain less than 10% of movements in aggregate activity. We then estimate key structural parameters of a three‐sector business‐cycle model country by country and find a disconnect between the importance assigned to terms‐of‐trade shocks by theoretical and SVAR models.     

是什么导致了“污染天堂”:贸易还是FDI?——来自中国省级面板数据的证据

本文利用1999-2008年中国省级面板数据,应用新近发展的面板向量自回归模型,研究FDI、贸易和环境规制之间的互动关系。结果发现,污染天堂假说在中国基本成立,而导致中国环境压力增大的最主要因素,并非是外商直接投资,而是自由贸易。这个结论有力地支持了西方污染中国的命题。进一步分析还发现,环境规制与对外贸易存在非对称的互动关系:环境规制对对外贸易存在显著负面影响,证实了污染天堂效应的存在;但反过来,对外贸易在一定程度上有利于环境治理努力的增强。最后本文选择多种环境规制指标和划分东、中、西部子样本的方法进行稳健性分析,证实了结论的可靠性。     

Why does private consumption rise after a government spending shock?

Abstract.  Some recent empirical evidence suggests that private consumption is crowded-in by government spending. This outcome violates neoclassical macroeconomic theory, according to which the negative wealth effect brought about by a rise in public expenditure should decrease consumption. In this paper, we develop a simple real business cycle model where preferences depend on private and public spending, and households are habit forming. The model is estimated by the maximum-likelihood method using U.S. data. Estimation results indicate a strong Edgeworth complementarity between private and public spending. This feature enables the model to generate a positive response of consumption following a government spending shock. In addition, the impulse-response functions generated by the estimated model are generally consistent with those obtained from a benchmark vector autoregression.     

Combination forecasts of tourism demand with machine learning models

The main objective of this study is to analyse whether the combination of regional predictions generated with machine learning (ML) models leads to improved forecast accuracy. With this aim, we construct one set of forecasts by estimating models on the aggregate series, another set by using the same models to forecast the individual series prior to aggregation, and then we compare the accuracy of both approaches. We use three ML techniques: support vector regression, Gaussian process regression and neural network models. We use an autoregressive moving average model as a benchmark. We find that ML methods improve their forecasting performance with respect to the benchmark as forecast horizons increase, suggesting the suitability of these techniques for mid- and long-term forecasting. In spite of the fact that the disaggregated approach yields more accurate predictions, the improvement over the benchmark occurs for shorter forecast horizons with the direct approach.     

Has US inflation really become harder to forecast?

Since the mid-1980s, Phillips curve forecasts of US inflation have been inferior to those of a conventional causal autoregression. However, little change in forecast accuracy is detected against the benchmark of a noncausal autoregression, more accurately characterizing US inflation dynamics.     

Oil Prices, Terms of Trade Shocks, and Macroeconomic Fluctuations in Saudi Arabia

The article investigates the sources of macroeconomic fluctuations in Saudi Arabia using structural vector autoregression methods and pays particular attention to oil prices and changes in terms of trade. Using a macroeconomic model tailored to the Saudi Arabian economy, the authors identify terms of trade, supply, balance of payments, aggregate demand, and monetary shocks. The results show that the Saudi Arabian price level, real exchange rate, and to a lesser extent output is vulnerable to terms of trade shocks. Moreover, Saudi Arabian terms of trade are driven by output, trade balance, and aggregate demand shocks. To stabilize output and the real exchange rate, Saudi Arabia ought to continue diversifying its production base and aim for a stable nominal oil price. (JEL E32 , Q43 , C22 )     

DATA REVISIONS AND FORECASTING ACCURACY: AN ECONOMETRIC ANALYSIS BASED ON PRELIMINARY AND REVISED NATIONAL ACCOUNTING ESTIMATES*

This paper entails an investigation of the effects of data revisions on forecasting accuracy, through use of preliminary and revised national accounting data compiled by the United Nations. A small model was estimated for each of fourteen countries and ex post“forecasts” generated for each country and each year of the period 1957–1964, using first preliminary and then revised data. A prior analysis of the data revisions indicated a strong and widespread tendency for the preliminary estimates to understate both levels and year-to-year changes. This is consistent with the findings of other studies. Two sets of forecasts obtained from the reduced form of the model were considered in relation to “actual” levels and changes, obtained from the revised data, and also in relation to each other. A strong downward bias was observed in the forecasts of levels based on preliminary data, and a weaker one in the forecasts of changes. The forecast discrepancies for different variables were found to be significantly correlated. The results suggest that a tendency toward understatement in preliminary data may account in part for the general tendency toward understatement in forecasts noted in other studies.     

Modelling the implied volatility surface based on Shanghai 50ETF options

We develop a dynamic factor model to forecast the implied volatility surface (IVS) of Shanghai Stock Exchange 50ETF options. Based on the assumption that dynamic change in IVS is mean-reverting and Markovian, we use a state space model to capture the dynamics of IVS, and set the latent factors to be the Ornstein–Uhlenbeck processes. We obtain the optimal estimations of parameters using the Kalman filter algorithm. Empirical results show that our model performs better than the traditional IVS model in terms of fitting ability and prediction performance.     

NORTH AMERICAN FREE TRADE AGREEMENT BILATERAL TRADE EFFECTS

This study evaluates the effects of the North American Free Trade Agreement (NAFTA) on bilateral trade between the United States and Canada and between the United States and Mexico. Trade flow estimates are from a vector autoregression (VAR) model. The VAR methodology allows modeling bilateral trade in a flexible manner that incorporates both the interaction between different variables and the dynamics of trade, output, prices, and the exchange rate. After testing the outside sample forecasting ability of the models, the study produces dynamic forecasts of bilateral trade. It then compares forecasts incorporating the effects of the NAFTA with baseline forecasts. The results suggest expanded trade for all three countries and an improvement in the U.S. trade position with both Canada and Mexico.     

FORECASTING THE YIELD CURVE USING PRIORS FROM NO‐ARBITRAGE AFFINE TERM STRUCTURE MODELS*

I propose a strategy for forecasting the term structure of interest rates that may produce significant gains in predictive accuracy. The key idea is to use the restrictions implied by Gaussian, no‐arbitrage, affine term structure models on a vector autoregression as prior information instead of imposing the restrictions dogmatically. This allows us to account for possible model misspecification. We use the proposed method to forecast a system of five U.S. yields up to 12 months ahead, and we find it provides significant gains in forecast accuracy.     

Does the vector error correction model perform better than others in forecasting stock price? An application of residual income valuation theory

This paper employs a multi-equation model approach to consider three statistic problems (heteroskedasticity, endogeneity and persistency), which are sources of bias and inefficiency in the predictive regression models. This paper applied the residual income valuation model (RIM) proposed by Ohlson (1995) to forecast stock prices for Taiwan three sectors. We compare relative forecasting accuracy of vector error correction model (VECM) with the vector autoregressive model (VAR) as well as OLS and RW models used in the prior studies. We conduct out-of-sample forecasting and employ two instruments to assess forecasting performance. Our empirical results suggest that the VECM statistically outperforms other three models in forecasting stock prices. When forecasting horizons extend longer, VECM produces smaller forecasting errors and performs substantially better than VAR, suggesting that the ability of VECM to improve VAR forecast accuracy is stronger with longer horizons. These findings imply that an error correction term (ECT) of the VECM contributes to improving forecast accuracy of stock prices. Our economic significance analyses and robustness tests for different data frequency are in favor of the superiority of VECM estimator.     

NEUTRAL TECHNOLOGY SHOCKS AND THE DYNAMICS OF LABOR INPUT: RESULTS FROM AN AGNOSTIC IDENTIFICATION*

This article studies the dynamic response of labor input to neutral technology shocks. It uses benchmark dynamic, stochastic, general equilibrium models enriched with labor market search and matching frictions and investment‐specific technological progress that enables a new, agnostic, identification scheme based on sign restrictions on a structural vector autoregression (SVAR). The estimation supports an increase of labor input in response to neutral technology shocks. This finding is robust across different perturbations of the SVAR model.     

Quantile forecast combination using stochastic dominance

This paper derives optimal forecast combinations based on stochastic dominance efficiency (SDE) analysis with differential forecast weights for different quantiles of forecast error distribution. For the optimal forecast combination, SDE will minimize the cumulative density functions of the levels of loss at different quantiles of the forecast error distribution by combining different time-series model-based forecasts. Using two exchange rate series on weekly data for the Japanese yen/US dollar and US dollar/Great Britain pound, we find that the optimal forecast combinations with SDE weights perform better than different forecast selection and combination methods for the majority of the cases at different quantiles of the error distribution. However, there are also some very few cases where some other forecast selection and combination model performs equally well at some quantiles of the forecast error distribution. Different forecasting period and quadratic loss function are used to obtain optimal forecast combinations, and results are robust to these choices. The out-of-sample performance of the SDE forecast combinations is also better than that of the other forecast selection and combination models we considered.     

Forecasting with large datasets: compressing information before,during or after the estimation?

We study the forecasting performance of three alternative large data forecasting approaches. These three approaches handle the dimensionality problem evoked by a large dataset by compressing its informational content, yet at different stages of the forecasting process. We consider different factor models, a large Bayesian vector autoregression and model averaging techniques, where the data compression takes place before, during and after the estimation of the respective forecasting models. We use a quarterly dataset for Germany that consists of 123 variables and find that overall the large Bayesian vector autoregression and the Bayesian factor augmented vector autoregression provide the most precise forecasts for a set of 11 core macroeconomic variables. Further, we find that the performance of these two models is very robust to the exact specification of the forecasting model.     

Reuters Sentiment and Stock Returns

Sentiment from more than 3.6 million Reuters news articles is tested in a vector autoregression model framework on its ability to forecast returns of the Dow Jones Industrial Average stock index. We show that Reuters sentiment can explain and predict changes in stock returns better than macroeconomic factors. We further find that negative Reuters sentiment has more predictive power than positive Reuters sentiment. Trading strategies with Reuters sentiment achieve significant outperformance with high success rates as well as high Sharpe ratios.     

Impacts of RMB devaluation on China’s trade balances: a time-varying SVAR approach

ABSTRACT

In this article, I empirically examine time-varying effects of real renminbi (RMB) devaluation on China’s trade balances. To this end, I estimate a time-varying structural vector autoregression model with monthly data. Results demonstrated that effects of real RMB devaluation on China’s trade balances are time varying. In a few months after devaluation of RMB, trade balances worsen as predicted in J-curve theory in most sample periods. However, subsequent improvements of trade balances predicted in J-curve theory appear only in certain periods, particularly in 2000s. Finally, devaluation of RMB positively affects China’s GDP and OECD industrial production, although the size of effects varies across sample periods. Joining WTO, the global financial crisis and endogenous feedbacks induced by price effects seem to be important to understand these time-varying patterns.     

FORECASTING WOMEN,INFANTS, AND CHILDREN CASELOADS: A COMPARISON OF VECTOR AUTOREGRESSION AND AUTOREGRESSIVE INTEGRATED MOVING AVERAGE APPROACHES

Under the Special Supplemental Nutrition Program for Women, Infants, and Children (WIC), each state receives a fixed federal grant for the operation of WIC in the upcoming federal fiscal year. Accurate forecasting is vital because states have to bear the expenses of any underestimation of WIC expenditures. Using monthly data from 1997 through 2005, this paper examined the performance of two competing models, autoregressive integrated moving average (ARIMA) and vector autoregression (VAR), in forecasting New York WIC caseloads for women, infants, and children. VAR model predicted over $120,000 less per month in forecast errors in comparison to the ARIMA model. (JEL H7, C5)     

Forecasting the Government Bond Term Structure in Australia

In this paper, we evaluate the performance of the dynamic Nielsen and Siegel interest rate model in forecasting Australian government bond yields. We compare a two‐stage OLS estimation procedure to a more powerful and robust state‐space framework estimated via a Kalman filter. We show that the one‐step approach generates smaller forecast errors than the two‐step procedure or a benchmark random walk model when forecasting the Australian government term structure across various horizons.