D-GMDH: A novel inductive modelling approach in the forecasting of the industrial economy

This work proposes a new forecasting model to analyse the economic development of Sichuan province of China. The model, which introduces the concept of diversity, is based on an improvement of the -GMDH algorithm. The new method, called D-GMDH, is compared with two ensemble approaches which are introduced by Dutta (2009), and D-GMDH is better than the two approaches in forecasting accuracy. D-GMDH is also applied to forecast the industrial added value of the Sichuan province. The obtained results are compared with those of the traditional GMDH model, GMDH combination model and the widely used ARMA model. The results show that D-GMDH has good prediction accuracy and is an effective means for economic forecasting when data is contaminated by noise.     

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.     

Synthesis of qualitative and quantitative approaches to long-range forecasting

Integration of qualitative and quantitative forecasting approaches is usually attempted and appraised in the context of improving forecast accuracy. This paper examines the feasibility and potential merits of integrating these two approaches to produce a useful range, rather than to improve the accuracy, of long-range forecasting. Two approaches of synthesis, one beginning with qualitative future scenarios and the other beginning with quantitative econometric models, are described, illustrated, and compared.     

Forecasting MBT technologies using DEA and LR

First, this paper explores Main Battle Tank (MBT) data set with different statistical methods in order to decide the most appropriate variables as reliable yardsticks in applying technology forecasting (TF) using data envelopment analysis (TFDEA) technique. It then applies TF using DEA method to forecast MBT technologies. This article attempts to predict technology development year of MBT commercialised from 1941 to 1994. This article presents the processes of TFDEA in detail and identifies some issues to search for appropriate input and output variables to forecast MBT technologies. The purpose of this study is to address some issues and identify an appropriate data to predict future trends of MBT technologies when using TFDEA and multiple linear regression tools. Finally, the study provides an understanding of the technological advances being sought in MBT technologies and information for use in making decisions regarding development strategy.     

Interest rate assumptions and predictive accuracy of central bank forecasts

The interest rate assumptions for macroeconomic forecasts differ among central banks. Common approaches are given by the assumptions that interest rates remain constant over the forecast horizon, follow a path as expected by market participants or follow a path as expected by the central bank itself. Theoretical papers such as Svensson (The instrument-rate projection under inflation targeting: the Norwegian example. Centre for European Policy Studies Working Paper (127), 2006) and Galí (J Monet Econ 58:537–550, 2011) suggest an accuracy ranking for these forecasts, from employing central bank expectations yielding the highest forecast accuracy to conditioning on constant interest rates yielding the lowest. Yet, when investigating the predictive accuracy of the Bank of England’s and the Banco Central do Brasil’s forecasts for interest rates, inflation and output growth, we hardly find any significant differences between forecasts based on the different interest rate paths. Our results suggest that the choice of the interest rate assumption appears to be of minor relevance empirically.     

Forecasting the US real house price index: Structural and non-structural models with and without fundamentals

We employ a 10-variable dynamic structural general equilibrium model to forecast the US real house price index as well as its downturn in 2006:Q2. We also examine various Bayesian and classical time-series models in our forecasting exercise to compare to the dynamic stochastic general equilibrium model, estimated using Bayesian methods. In addition to standard vector-autoregressive and Bayesian vector autoregressive models, we also include the information content of either 10 or 120 quarterly series in some models to capture the influence of fundamentals. We consider two approaches for including information from large data sets — extracting common factors (principle components) in factor-augmented vector autoregressive or Bayesian factor-augmented vector autoregressive models as well as Bayesian shrinkage in a large-scale Bayesian vector autoregressive model. We compare the out-of-sample forecast performance of the alternative models, using the average root mean squared error for the forecasts. We find that the small-scale Bayesian-shrinkage model (10 variables) outperforms the other models, including the large-scale Bayesian-shrinkage model (120 variables). In addition, when we use simple average forecast combinations, the combination forecast using the 10 best atheoretical models produces the minimum RMSEs compared to each of the individual models, followed closely by the combination forecast using the 10 atheoretical models and the DSGE model. Finally, we use each model to forecast the downturn point in 2006:Q2, using the estimated model through 2005:Q2. Only the dynamic stochastic general equilibrium model actually forecasts a downturn with any accuracy, suggesting that forward-looking microfounded dynamic stochastic general equilibrium models of the housing market may prove crucial in forecasting turning points.     

A new multiscale decomposition ensemble approach for forecasting exchange rates

Due to the high complexity and strong nonlinearity nature of foreign exchange rates, how to forecast foreign exchange rate accurately is regarded as a challenging research topic. Therefore, developing highly accurate forecasting method is of great significance to investors and policy makers. A new multiscale decomposition ensemble approach to forecast foreign exchange rates is proposed in this paper. In the approach, the variational mode decomposition (VMD) method is utilized to divide foreign exchange rates into a finite number of subcomponents; the support vector neural network (SVNN) technique is used to model and forecast each subcomponent respectively; another SVNN technique is utilized to integrate the forecasting results of each subcomponent to generate the final forecast results. To verify the superiority of the proposed approach, four major exchange rates were chosen for model comparison and evaluation. The experimental results indicate that our proposed VMD-SVNN-SVNN multiscale decomposition ensemble approach outperforms some other benchmarks in terms of forecasting accuracy and statistical tests. This demonstrates that our proposed VMD-SVNN-SVNN multiscale decomposition ensemble approach is promising for forecasting foreign exchange rates.     

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.     

Forecasting house prices using dynamic model averaging approach: Evidence from China

Forecasting house price has been of great interests for macroeconomists, policy makers and investors in recent years. To improve the forecasting accuracy, this paper introduces a dynamic model averaging (DMA) method to forecast the growth rate of house prices in 30 major Chinese cities. The advantage of DMA is that this method allows both the sets of predictors (forecasting models) as well as their coefficients to change over time. Both recursive and rolling forecasting modes are applied to compare the performance of DMA with other traditional forecasting models. Furthermore, a model confidence set (MCS) test is used to statistically evaluate the forecasting efficiency of different models. The empirical results reveal that DMA generally outperforms other models, such as Bayesian model averaging (BMA), information-theoretic model averaging (ITMA) and equal-weighted averaging (EW), in both recursive and rolling forecasting modes. In addition, in recent years it is found that the Google search index, instead of fundamental macroeconomic or monetary indicators, has developed greater predictive power for house price in China.     

The public value of improving a weather forecasting system in Korea: a choice experiment study

The Korean government plans to improve the quality of its weather forecasting system in order to increase its public utility. The benefits arising from the implementation of this plan should be measured. To this end, this study applies a choice experiment to four attributes: the update frequency of both short- and medium-range forecasts, and the accuracy of both. A survey of 1000 randomly selected households was undertaken in Korea. In the study results, the marginal willingness-to-pays, respectively, for one more update of the short-range forecast per day, for a 1% increase in the accuracy of the short-range forecast, for changing the update frequency of the medium-range forecast from once a day (reference level) to twice a day, and for a 1% increase in the accuracy of the medium-range forecast as a result of improving the weather forecast service were estimated to be KRW 499.3 (USD 0.45), 108.3 (0.10), 346.5 (0.31), and 80.9 (0.07) per household per month. The findings can provide policy-makers with useful information for both evaluating and planning improvements in the weather forecasting system.     

分析师预测信息来源的新发现:会计师事务所

文章通过检验券商与公司聘任同一家会计师事务所对券商旗下分析师预测行为的影响,研究了证券分析师的预测信息是否可能来源于会计师事务所.研究发现:(1)分析师更愿意跟踪与所属券商聘任同一家会计师事务所的上市公司,对其发布的盈余预测更准确,也更倾向于额外发布现金流预测.(2)券商与公司从非同聘会计师事务所变更为同聘会计师事务所,其分析师预测的准确性提高,反之则降低.分析师对与所属券商聘任同一家会计师事务所的公司在年报披露前最后一次盈余预测的准确性有更大幅度的提高.(3)同聘会计师事务所对预测准确性的促进作用集中在分析师跟踪少和收入变化大的公司以及非明星分析师发布的预测.文章的研究有助于理解同聘会计师事务所对分析师预测行为的影响,也有助于资本市场上的投资者更好地利用同聘会计师事务所的券商旗下分析师发布的预测报告,拓展了分析师预测信息来源的研究.     

Cross-Country Evidence on the Ability of the Nominal Interest Rate to Predict Inflation

This paper examines the viability of using short-term interest rates to forecast inflation as implied by the Fisher hypothesis. A major problem with this approach lies in the implicit assumptions that the real interest rate is constant and that the relationship between inflation and interest rate does not change over time. We demonstrate, using quarterly data for four OECD countries, that by relaxing these assumptions and allowing for seasonality in the inflation rate it is possible to obtain a model with a high degree of forecasting accuracy and efficiency.
JEL Classification Numbers: C22, C52, E31.     

A novel time-series model based on empirical mode decomposition for forecasting TAIEX

Stock price prediction is regarded as a challenging task of the financial time series prediction process. Time series models have successfully solved prediction problems in many domains, including the stock market. Unfortunately, there are two major drawbacks in stock market by time-series model: (1) some models cannot be applied to the datasets that do not follow the statistical assumptions; and (2) most time-series models which use stock data with many noises involutedly (caused by changes in market conditions and environments) would reduce the forecasting performance. For solving the above problems and promoting the forecasting performance of time-series models, this paper proposes a hybrid time-series support vector regression (SVR) model based on empirical mode decomposition (EMD) to forecast stock price for Taiwan stock exchange capitalization weighted stock index (TAIEX). In order to evaluate the forecasting performances, the proposed model is compared with autoregressive (AR) model and SVR model. The experimental results show that the proposed model is superior to the listing models in terms of root mean squared error (RMSE). And the more fluctuation year (2000–2001) occurs, the better accuracy of proposed model will be obtained.     

Imperfect competition and the modelling of expectations in macroeconomics

We show that when a model of the macroeconomy is based on imperfect, rather than perfect, competition, this may increase the problem of how to model agents’ expectations. We provide a simple example using an overlapping-generations economy with the potential for unemployment. Under certain assumptions about how consumers migrate between locations between the first and second periods of their lives, this extra issue regarding expectations arises. Imperfect competition may increase agents’ forecasting difficulties because they have to forecast not only future equilibrium prices, but also future out-of-equilibrium prices, and by definition the latter are never actually observed.

An international comparison of analysis' earnings forecast accuracy

A major task of financial analysts working for stockbrokers and investment firms is to forecast future earnings of listed companies. The usefulness of their work crucially depends on the accuracy of the forecasts. A great many studies have examined the accuracy, bias, and other characteristics of profit forecasts made in the U.S. In contrast, however, there is very little research on forecasting accuracy in other countries despite the increasingly global nature of investing. This paper examines the accuracy of corporate earnings forecasts in 34 different countries. In addition, a model is developed that seeks to explain differences across companies and countries. The findings show that eight countries have better forecast accuracy than the U.S. This cross-sectional model shows that with the inherent difficulty in forecasting for a specific company (proxied by the change in its earnings), risk and the number of analysts following the stock are the major factors in explaining earnings forecast accuracy.     

A patent-based cross impact analysis for quantitative estimation of technological impact: The case of information and communication technology

The increasing syntheses and interactions between various technologies increase the usefulness of cross impact analysis (CIA) as a method for forecasting and analyzing them. Conventional CIA depends on an expert's qualitative judgment or intuition and thus it is difficult to evaluate quantitatively the impact of one technology on another. In this study, we employ patent analysis in CIA to examine such impacts between technologies based on multiple patent classifications. Patent information is used for facilitating quantitative and systematic approach in CIA. The distinctive feature and main contribution of the proposed approach is the overcoming of the limitations of conventional CIA, by employing conditional probabilities based on the patent information. The classification of patents, particularly the multiple classifications, is used to evaluate the relationships between technologies. As an illustration, a patent-based CIA with information and communication technologies (ICTs) was conducted. Firstly, the patent-based cross impact among ICTs was calculated. Secondly, the technology pairs were classified based on the cross impact score between ICTs. Thirdly, a cross impact network was constructed to identify the complex relation among ICTs. Finally, the changes in cross impact scores between technologies over time were analyzed. The results of this research are expected to help practitioners to forecast future trends and to develop better R&D strategies.     

On a probabilistic model of forecasting

One of the methods of studying complex objects is the construction of a mathematical model, containing such information about the object that is necessary to solve a definite problem connected with it.Mathematical modeling, based on the construction of models of various kinds can be used in forecasting. Let a forecasting object A(X) be described by vector X = (X₁, X₂,…,X_n) whose coordinates are parameters characterizing this object. The work presents a probabilistic model of forecasting and gives the example of a forecast of the object described by a set two parameters.     

Comparative accuracy of forecasts of inflation: a Canadian study

An inaccurate forecast of inflation is costlier to economic agents when the inflation rate is high and volatile. In this situation, the use of more sophisticated and information-oriented forecasting models become economically efficient. We test this hypothesis by analysing the forecasting accuracy of vector auto-regressive (VAR), auto-regressive integrated moving average (ARIMA) and static expectation models. We use Canadian data and divide the post-sample forecasting period into four sub-periods, based on high/low and volatile/stable inflation. Prediction errors are compared for both short-term and long-term forecasts. Finally, the paper proposes a portfolio approach for obtaining a more accurate forecast of inflation.     

Projecting soviet energy requirements using a vintage capital model

A Solow-type vintage capital model with capital utilization is used to estimate the long-term average embedded rate of improvement in energy efficiency of Soviet fixed-capital commissionings: 2.6% per year. The results are combined with planned Soviet capital stock growth to forecast cap Soviet energy requirements through 1990. The implied Soviet capital utilization goals call for greater energy savings than have been achieved in the past. The results also identify an upper bound for rates of hidden inflation contained in Soviet capital commissionings data. It appears an estimate such as S. Cohn's 1% (Soviet Stud. 33, 2:269–299, 1981) is far more likely than A. Nove's 6 to 7% (Soviet Stud. 33, 1:142–145, 198 1).     

Forecasting Canadian mortgage rates

Mortgage rates are one of the important drivers of the housing market. While there is a literature looking at the pass-through effect from Central Bank rates to mortgage rates, there is less known about how useful Central Bank rates are for forecasting mortgage rates. This article uses a selection of models (ARIMA, ARIMAX, BATS, state space error, trend seasonal (ETS), Holt Winter, random walk, simple exponential smoothing (SES), OLS and VAR) to forecast Canadian 5-year conventional mortgage rates. Based on RMSE, regression-based approaches like ARIMAX or OLS that use Central Bank rates to forecast mortgage rates are preferred when it comes to forecasting Canadian mortgage rates 6 or 12 months into the future, respectively.