Improving forecast accuracy of financial vulnerability: PLS factor model approach

We present a factor augmented forecasting model for assessing the financial vulnerability in Korea. Dynamic factor models often extract latent common factors from a large panel of time series data via the method of the principal components (PC). Instead, we employ the partial least squares (PLS) method that estimates target specific common factors, utilizing covariances between predictors and the target variable. Applying PLS to 198 monthly frequency macroeconomic time series variables and the Bank of Korea's Financial Stress Index (KFSTI), our PLS factor augmented forecasting models consistently outperformed the random walk benchmark model in out-of-sample prediction exercises in all forecast horizons we considered. Our models also outperformed the autoregressive benchmark model in short-term forecast horizons. We expect our models would provide useful early warning signs of the emergence of systemic risks in Korea's financial markets.     

Do disaggregated CPI data improve the accuracy of inflation forecasts?

In this paper, we evaluate the role of using consumer price index (CPI) disaggregated data to improve the accuracy of inflation forecasts. Our forecasting approach is based on extracting the factors from the subcomponents of the CPI at the highest degree of disaggregation. The data set contains 54 macroeconomic series and 243 CPI subcomponents from 1992 to 2009 for Mexico. We find that the factor models that include disaggregated data outperform the benchmark autoregressive model and the factor models containing alternative groups of macroeconomic variables. We provide evidence that using disaggregated price data improves forecasting performance. The forecasts of the factor models that extract the information from the CPI disaggregated data are as accurate as the forecasts from the survey of experts.     

The possibility of using search traffic information to explore consumer product attitudes and forecast consumer preference

In recent years, many researchers have devoted their attention to using search traffic information gathered from Google Insights to carry out consumer attitude research. The purpose of this study is to assess the effectiveness of using search traffic information to analyze actual consumer attitudes regarding a product. By comparing the results of conventional survey-based attitude research with the results of search traffic information, this study reveals that search traffic information indicates consumers' level of interest regarding a product, the product attributes that they are considering, and the importance of each attribute to them. Also, it demonstrates the potential benefits of search traffic analysis, which can be useful for forecasting consumer preferences regarding products. Focusing on the Prius, a hybrid car, this study shows that search traffic information serves as an accurate indicator of consumer attitudes, and even succeeds in identifying consumers' hidden attitudes toward the Prius, which can be explained by cognitive dissonance theory. Finally, this study utilizes search traffic information to forecast changes in consumer attitudes and to develop an econometric model of consumer demand for the Prius by incorporating environmental variables such as the WTI (West Texas Intermediate) price. This study concludes that search traffic information offers new potential advantages, in that it not only overcomes the limitations imposed by the high cost of conducting surveys, in terms of money and time, but also helps to reduce the distortions caused by conscious or unconscious errors committed by survey respondents.     

MONTHLY GDP FORECASTING USING BRIDGE MODELS: APPLICATION FOR THE FRENCH ECONOMY

This paper presents a model to predict French gross domestic product (GDP) quarterly growth rate. The model is designed to be used on a monthly basis by integrating monthly economic information through bridge models, for both supply and demand sides, allowing thus economic interpretations. For each GDP component, bridge equations are specified by using a general‐to‐specific approach implemented in an automated way by Hoover and Perez and improved by Krolzig and Hendry. This approach allows to select explanatory variables among a large data set of hard and soft data. A rolling forecast study is carried out to assess the forecasting performance in the prediction of aggregated GDP, by taking publication lags into account in order to run pseudo real‐time forecasts. It turns out that the model outperforms benchmark models. The results show that changing the set of equations over the quarter is superior to keeping the same equations over time. In addition, GDP growth seems to be more precisely predicted from a supply‐side approach rather than a demand‐side approach.     

Demand forecasting for multigenerational products combining discrete choice and dynamics of diffusion under technological trajectories

The discrete choice model generally captures consumers' valuation of the product's quality within the framework of a cross-sectional analysis, while the diffusion model captures the dynamics of demand within the framework of a time-series analysis. We propose an adjusted discrete choice model that incorporates the choice behavior of the consumer into the dynamics of product diffusion. In addition, a new estimation structure is proposed, within the framework of the time-series analysis, which enables the estimation of the discrete choice model on market-level data to be performed in such a way as to avoid the problem of price endogeneity and to obtain greater flexibility in forecasting demand. As an empirical application, the suggested model is applied to the case of the worldwide DRAM (dynamic random access memory) market. In forecasting future demand of DRAM generations, we integrate Moore's law and learning by doing to reflect the future technological trajectories of DRAM innovations, as well as consumers' consumption trends to reflect the dynamics of demand environments. As a result, the suggested model shows better performance in explaining the diffusion of new-generation product with limited number of data observations.     

基于ARIMA与GARCH模型的国际油价预测比较分析

在分析影响油价波动因素的基础上,利用1986年1月至2010年12月的WTI国际原油价格月度数据,分别建立ARIMA和GARCH模型对油价进行预测。并通过对2011年1月至2012年4月WTI原油价格进行外推预测,检验模型的预测效果。比较分析发现,在短期预测中,ARIMA和GARCH模型对油价的预测均比较准确,但当油价由于受到重大事件的影响而有较大波动时,模型的预测精度下降;在长期预测中,GARCH模型的预测效果优于ARIMA模型;整体来看,GARCH模型预测的精度高于ARIMA模型。因此,在国际油价预测中,用GARCH模型是比较合适的。     

Estimating and testing the compensated double-log demand model

In spite of the proliferation of flexible functional forms for consumer demand systems, the double-log demand model continues to be popular, especially in applied work calling for single-equation models. It is usually estimated in uncompensated form. It can also be estimated in compensated form, by deflating the income variable alone using Stone's price index. The compensated form has the same right-hand side as a single-equation version of the popular linear approximation to the Almost Ideal demand model, facilitating the construction of a test for choosing between the two alternatives. This paper demonstrates these results, develops the specification test, and illustrates its application using US meat consumption data. Simulations suggest that the test is well-behaved with good power in typical applications.     

Forecasting future demand for large-screen television sets using conjoint analysis with diffusion model

The methodological framework proposed in this paper addresses two limitations of the basic Bass diffusion model: that it does not reflect competition among products nor does it forecast demand for products that do not exist in the marketplace. The model consists of four steps. First, to investigate consumer preferences for product attributes, we use conjoint analysis to estimate the utility function of consumers. Next we estimate the dynamic price function of each competing product to reflect technological changes and the evolving market environment. Then we derive dynamic utility function by combining the static utility function and the price function. Finally, we forecast the sales of each product using estimated market share and sales data for each period, which are derived from the dynamic utility function and from the Bass diffusion model, respectively.We apply this model to South Korea's market for large-screen televisions. The results show that (1) consumers are sensitive to picture resolution and cost and (2) in the near future, should the market see the introduction of liquid crystal display (LCD) TVs with screens larger than 50 inches, the high resolution and steep price drop of LCD will lead LCD TVs to capture a larger market share than TVs with other display types. Finally, our results show that TVs with 40-inch screens are preferred over TVs with larger screens.     

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.     

Forecasting US consumer price index: does nonlinearity matter?

The objective of this article is to predict, both in sample and out of sample, the consumer price index (CPI) of the US economy based on monthly data covering the period of 1980:1–2013:12, using a variety of linear (random walk (RW), autoregressive (AR) and seasonal autoregressive integrated moving average (SARIMA)) and nonlinear (artificial neural network (ANN) and genetic programming (GP)) univariate models. Our results show that, while the SARIMA model is superior relative to other linear and nonlinear models, as it tends to produce smaller forecast errors; statistically, these forecasting gains are not significant relative to higher-order AR and nonlinear models, though simple benchmarks like the RW and AR(1) models are statistically outperformed. Overall, we show that in terms of forecasting the US CPI, accounting for nonlinearity does not necessarily provide us with any statistical gains.     

Forecasting UK consumer price inflation using inflation forecasts

The inflation rate is a key economic indicator for which forecasters are constantly seeking to improve the accuracy of predictions, so as to enable better macroeconomic decision making. Presented in this paper is a novel approach which seeks to exploit auxiliary information contained within inflation forecasts for developing a new and improved forecast for inflation by modeling with Multivariate Singular Spectrum Analysis (MSSA). Unlike other forecast combination techniques, the key feature of the proposed approach is its use of forecasts, i.e. data into the future, within the modeling process and extracting auxiliary information for generating a new and improved forecast. We consider real data on consumer price inflation in UK, obtained via the Office for National Statistics. A variety of parametric and nonparametric models are then used to generate univariate forecasts of inflation. Thereafter, the best univariate forecast is considered as auxiliary information within the MSSA model alongside historical data for UK consumer price inflation, and a new multivariate forecast is generated. We find compelling evidence which shows the benefits of the proposed approach at generating more accurate medium to long term inflation forecasts for UK in relation to the competing models. Finally, through the discussion, we also consider Google Trends forecasts for inflation within the proposed framework.     

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.     

World steel production: A new monthly indicator of global real economic activity

This paper proposes world steel production as an indicator of global real economic activity. World steel production data is published with only a one-month delay, thereby providing timely information for world real GDP forecasters. We find that world steel production and Lutz Kilian's (2009) index of global real economic activity generate large gains in forecasting world real GDP, relative to an autoregressive benchmark. A forecast combination of world steel production, Kilian's (2009) index of global real economic activity and an index of the industrial production of OECD countries plus six non-OECD emerging economies produces significant gains in forecasting world real GDP, relative to an autoregressive benchmark     

OWA-based ANFIS model for TAIEX forecasting

In stock market forecasting, high-order time-series models that use previous several periods of stock prices as forecast factors are more reasonable to provide a superior investment portfolio for investors than one-order time-series models using one previous period of stock prices. However, in forecasting processes, it is difficult to deal with high-order stock data, because it is hard to give a proper weight to each period of past stock price, reduce data dimensions without losing stock information, and produce a comprehensive forecasting result based on stock data with nonlinear relationships.Additionally, there are two more drawbacks to past time-series models: (1) some assumptions (Bollerslev, 1986; Engle, 1982) about stock variables are required for statistical methods, such as the autoregressive model (AR) and autoregressive moving average (ARMA); (2) numeric time-series models have been presented to deal with forecasting problems for stock markets, but they can not handle the nonlinear relationships within the stock prices.To address these shortcomings, this paper proposes a new time series model, which employs the ordered weighted averaging (OWA) operator to fuse high-order data into the aggregated values of single attributes, a fusion adaptive network-based fuzzy inference system (ANFIS) procedure, for forecasting stock price in Taiwanese stock markets.In verification, this paper employs a seven-year period of the TAIEX stock index, from 1997 to 2003, as experimental datasets and the root mean square error (RMSE) as evaluation criterion. The experimental results indicate that the proposed model is superior to the listing methods in terms of root mean squared error.     

Impacts of meat product recalls on consumer demand in the USA

The impact of meat product recall events on consumer demand (beef, pork, poultry, and other consumption goods) in the USA is tested empirically. Beef, pork, and poultry recall indices are constructed from both the Food Safety Inspection Service's meat recall events and from newspaper reports over the period 1982–1998. Following previous product recall studies, recall indices are incorporated as shift variables in consumers’ demand functions. Estimating an absolute price version of the Rotterdam demand model, findings indicate that Food Safety Inspection Service's meat recall events significantly impact demand, and newspaper reports do not. Moreover, although elasticities related to recall events are significant they are small in magnitude relative to price and income effects. Any favourable effects on the demands of meat substitutes for a recall are offset by a more general negative effect on meat demand. The general negative effect indicates a shift out of meat to non-meat consumption goods.     

Forecasting with preliminary data: a comparison of two methods

This study examines two alternate methods, a vector autoregression error correction model and a state space model, to forecast revised United States trade balance figures. Both these methods incorporate preliminary and revised trade data. The results obtained from these methods were compared to the benchmark forecasts generated by revised-data-only models. This Study finds that the state space model performs worse than the benchmark. The vector autoregression model performs better than the benchmark only in the one-step forecast. These results indicate that incorporating preliminary data may not be useful in forecasting the revised data.     

基于小波自适应支持向量机的短期电价预测

电价预测对于发电商、供电企业以及市场监管者都具有重要的意义。提出一种小波自适应支持向量机预测模型,先将电价时间序列作小波分解得到低频和高频分量,再采用自适应调整法,自动地为支持向量机选择较好的参数对电价小波分量逐一预测,最后通过小波重构得到电价最终预测结果。实例证明前述方法得到的预测精度高于BP、RBF、SVM等传统预测模型。     

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.     

Competition and confidentiality: Signaling quality in a duopoly when there is universal private information

We model non-cooperative signaling by two firms that compete over a continuum of consumers, assuming each consumer has private information about the intensity of her preferences for the firms' respective products and each firm has private information about its own product's quality. We characterize a symmetric separating equilibrium in which each firm's price reveals its respective product quality. We show that the equilibrium prices, the difference between those prices, the associated outputs, and profits are all increasing functions of the ex ante probability of high safety. If horizontal product differentiation is sufficiently great then equilibrium prices and profits are higher under incomplete information about quality than if quality were commonly known. Thus, while signaling imposes a distortionary loss on a monopolist using price to signal quality, duopolists may benefit from the distortion as it can reduce competition. Finally, average quality is lower since signaling quality redistributes demand towards low-quality firms.     

SWITCHING COST AND DEPOSIT DEMAND IN CHINA

This article develops and estimates a dynamic model of consumer demand for deposits in which banks provide differentiated products and product characteristics that evolve over time. The switching cost is 0.8% of the deposit's value, which leads the static model to bias the demand estimates. The dynamic model shows that the price elasticity over a long time horizon is larger than the same elasticity over a short time horizon. Counterfactual experiments with a dynamic monopoly show that reducing the switching cost has a comparable competitive effect on bank pricing as a result of reducing the dominant position of the monopoly.