Unobserved Components Model Research Articles

Unemployment flows, participation, and the natural rate of unemployment: Evidence from turkey

We use a parsimonious unobserved components model with flow rates to estimate a time-varying unemployment rate trend for Turkey. Our approach is grounded in the modern theory of labor market search. This trend estimate yields a level that the unemployment rate would converge to in the absence of cyclical shocks that move different flow rates away from their underlying trends. By the end of 2018, the unemployment rate trend for Turkey stood at 12 percent, gradually increasing since the end of the global recession. The key drivers of the unemployment rate trend have been rising separation and unemployment exit rate trends in the first decade of our sample period. Since 2012, these trends have stagnated somewhat, but the labor force participation rate has continued trending up.

Nowcasting the Spanish Economy Using Very High Frequency Tax Data

We present a short-term forecasting model based on tax data. The model combines daily information from the Immediate Supply of Information System for VAT declaration forms, with monthly indicators derived from tax data. The model uses the GDP as a macroeconomic synthesis. The model combines signal extraction and forecasting at the daily frequency, by means of an unobserved components model, with a mixed frequency (monthly-quarterly) dynamic factor analysis for GDP now-casting and forecasting. The daily information, plus the flexibility and efficiency of the factor models, allows a permanently updated monitoring of the short-term economic conditions of the Spanish economy.

Temporal Disaggregation of Overlapping Noisy Quarterly Data: Estimation of Monthly Output from Uk Value-added Tax Data

SummaryThe paper derives monthly estimates of business sector output in the UK from rolling quarterly value-added tax based turnover data. The administrative nature of the value-added tax data implies that their use could ultimately yield a more precise and granular picture of output across the economy. However, they show two particular features which complicate their exploitation: they are overlapping and subject to substantial noise. This motivates our choice of a multivariate unobserved components model for filtering and disaggregating temporally the aggregate figures. After illustrating our method by using one industry as a case-study, we estimate monthly seasonally adjusted gross output figures for the 75 industries for which the data are available. Our results show material differences from the existing output profile.

Unobserved Components Model for Forecasting Sugarcane Production in Sri Lanka

Unobserved Components Model (UCM) is a structural time series model and it can decompose the response series into latent components, such as trend, cycle and seasonal effects and linear and nonlinear regression effects. The UCM combines the capabilities of Autoregressive Integrated Moving Average (ARIMA) model with interpretability of smoothing models. This study was carried out to forecast sugarcane production in Sri Lanka using UCM model. The best fitting model was selected based on Akaike information criterion (AIC) and Bayesian Information Criterion (BIC), followed by residual analysis. The selected model was used to make sample period forecasts (From 1979 - 2013) and post sample period forecasts (From 2014 to 2018). Forecasting accuracy of model was evaluated using the Mean Absolute Percentage Error (MAPE). Linear trend model (adj. R2=77 %) with zero variance slope and two cycles was selected as best among the tested UCM models for cane production data. MAPE was 10.56 % for sample period forecasts and 4.01 % for post-sample period forecasts. Predicted cane production for year 2019 was 813,888 ± 293,891 tons.

Modeling and predicting the patterns of seasonal rainfall in Tamil Nadu, India 1951–2017: an UCM approach

The main objective of the paper is to investigate the behavior, modeling, and predicting the patterns of the seasonal rainfall in Tamil Nadu using the unobserved components model (UCM) with the concealed components drift, seasonal, cyclical, and irregular. The India Meteorological Department (IMD) has provided the seasonal rainfall data of Tamil Nadu (TN), and the analysis has been done for winter, pre-monsoon, monsoon, and post-monsoon seasons. The UCM with deterministic components drift, cycle, and trigonometric seasonal and stochastic irregular component is selected for predicting the patterns of seasonal rainfall in TN from the parsimonious models based on the significant tests of free parameters and components, information criteria, and statistical fit. The maximum likelihood estimation is used for the model parameters, and the validity of the chosen model is resolved through normal and correlation diagnostics. The patterns of seasonal rainfall are predicted with 95% confidence interval during the years 2018–2020 in virtue of selected UCM. The identified UCM forecasts the winter rainfall around 11.12 mm in 2018, 8.57 mm in 2019, and 19.3 mm in 2020; pre-monsoon rainfall will be 132.72 mm in 2018, 116.74 mm in 2019, and 107.62 mm in 2020; monsoon rainfall will be 354.16 mm in 2018, 349.52 mm in 2019, and 334.65 mm in 2020; post-monsoon rainfall will be 440.84 mm in 2018, 454.68 mm in 2019, and 457.1 mm in 2020.

A non-parametric statistic for testing conditional heteroscedasticity for unobserved component models

When prediction intervals are constructed using unobserved component models (UCM), problems can arise due to the possible existence of components that may or may not be conditionally heteroscedastic. Accurate coverage depends on correctly identifying the source of the heteroscedasticity. Different proposals for testing heteroscedasticity have been applied to UCM; however, in most cases, these procedures are unable to identify the heteroscedastic component correctly. The main issue is that test statistics are affected by the presence of serial correlation, causing the distribution of the statistic under conditional homoscedasticity to remain unknown. We propose a nonparametric statistic for testing heteroscedasticity based on the well-known Wilcoxon's rank statistic. We study the asymptotic validation of the statistic and examine bootstrap procedures for approximating its finite sample distribution. Simulation results show an improvement in the size of the homoscedasticity tests and a power that is clearly comparable with the best alternative in the literature. We also apply the test on real inflation data. Looking for the presence of a conditionally heteroscedastic effect on the error terms, we arrive at conclusions that almost all cases are different than those given by the alternative test statistics presented in the literature.

An effcient exact Bayesian method For state space models with stochastic volatility

Abstract This paper introduces a Bayesian MCMC method, referred to as a marginalized mixture sampler, for state space models whose disturbances follow stochastic volatility processes. The marginalized mixture sampler is based on a mixture-normal approximation of the log-χ 2 distribution, but it is implemented without the need to simulate the mixture indicator variable. The key innovation is to use the filter ing scheme developed by Kim (Kim C.-J. 1994. “Dynamic Linear Models with Markov-Switching.” Journal of Econometrics 60: 1–22.) and the forward-filtering backward-sampling algorithm to generate a proposal series of the latent stochastic volatility process. The proposal series is then accepted according to the Metropolis-Hastings acceptance probability. The new sampler is examined within an unobserved component model and a time-varying parameter vector autoregressive model, and it reduces substantially the correlations between MCMC draws.

Unobserved component model with multi-interval input interventions: An application to crude oil production

In this paper, a multi-interval input intervention unobserved component model (MIII-UCM) is developed to model the crude oil production series. The focus of this study is on the development of a model that could be used to model and evaluate the effects of the military crackdown, presidential amnesty program, and Niger Delta avenger insurgency respectively. The monthly crude oil production dataset from the web database of the central bank of Nigeria (CBN) was used. The proposed UCM-intervention model with significant unobserved stochastic level, deterministic trigonometric seasonal, stochastic irregular term and pulse intervention functions based on the Akaike's information criteria (AIC), Schwartz's Bayesian Information Criteria (BIC), goodness-of-fit and significance tests captured the structural variability of the time series used in this study. The results showed that step intervention functions yield negative nonsignificant effects. Generally, the military crackdown on militant groups, the presidential amnesty program, and Niger Delta avenger insurgency presented a negative nonsignificant shift on crude oil production.

Bayesian analysis of output gap in Barbados

Abstract This article contributes to understanding the performance of various unobserved components (UC) models in fitting Barbados’ real GDP. Relying on recent UC models techniques, it finds support for the UC model that captures correlated disturbances, but not for the model that does not. The best model is the correlated UC model with two breaks: 1981 and 2008. Estimates show that the trend and cycle innovations are positively correlated. There is evidence suggesting that trend output growth has slowdown in the past decade. The Bayes factor result indicates that the GDP trend is better modeled as a stochastic process rather than a deterministic one.

Too Many Shocks Spoil the Interpretation

We show that when a model has more shocks than observed variables the estimated filtered and smoothed shocks will be correlated. This is despite no correlation being present in the data generating process. Additionally the estimated shock innovations may be autocorrelated. These correlations limit the relevance of impulse responses, which assume uncorrelated shocks, for interpreting the data. Excess shocks occur frequently, e.g. in Unobserved-Component (UC) models, filters, including Hodrick- Prescott (1997), and some Dynamic Stochastic General Equilibrium (DSGE) models. Using several UC models and an estimated DSGE model, Ireland (2011), we demonstrate that sizable correlations among the estimated shocks can result.

Too Many Shocks Spoil the Interpretation

We show that when a model has more shocks than observed variables the estimated filtered and smoothed shocks will be correlated. This is despite no correlation being present in the data generating process. Additionally the estimated shock innovations may be autocorrelated. These correlations limit the relevance of impulse responses, which assume uncorrelated shocks, for interpreting the data. Excess shocks occur frequently, e.g. in Unobserved-Component (UC) models, filters, including Hodrick-Prescott (1997), and some Dynamic Stochastic General Equilibrium (DSGE) models. Using several UC models and an estimated DSGE model, Ireland (2011), we demonstrate that sizable correlations among the estimated shocks can result.

Estimating Time-Varying Long-Run Growth Rate of Russian GDP in the ARX Model with Oil Prices

The paper estimates the path of trend growth rates for Russian GDP based on an autoregressive model with exogenous variables and with a time-varying parameter of trend growth, which, by assumption, is described by a random walk process. In conditions of a high dependence of the Russian economy on commodity exports, terms of trade are used as a control exogenous variable for GDP dynamics. For the purpose of econometric estimation, the ARX model is presented as an unobserved components model and estimated using the maximum likelihood method with the Kalman filter applied. It is shown that in the first half of the 2000s the trend growth rate was at 4%, which can be interpreted as recovery growth after a transformational recession. The higher growth rates actually achieved during this period are explained by the intensive growth of world oil prices. Later, the potential for recovery growth was exhausted, and after the crisis of 2008 the rates of trend growth were remaining at the level of 2% per year for a long period of time. However, following the 2014 crisis, trend growth rates began to decline steadily, and had reached about 1% per year by the beginning of 2019, which can be interpreted as the impact of sanctions and geopolitical uncertainty on the economic development of the Russian Federation. The results of an econometric analysis of the model on household consumption and investment data also suggest that the trend growth rate is approximately 1% per year at present.

Inflation and professional forecast dynamics: An evaluation of stickiness, persistence, and volatility

This paper studies the joint dynamics of U.S. inflation and a term structure of average inflation predictions taken from the Survey of Professional Forecasters (SPF). We estimate these joint dynamics by combining an unobserved components (UC) model of inflation and a sticky‐information forecast mechanism. The UC model decomposes inflation into trend and gap components, and innovations to trend and gap inflation are affected by stochastic volatility. A novelty of our model is to allow for time‐variation in inflation‐gap persistence as well as in the frequency of forecast updating under sticky information. The model is estimated with sequential Monte Carlo methods that include a particle learning filter and a Rao–Blackwellized particle smoother. Based on data from 1968 Q4 to 2018 Q3, estimates show that (i) longer horizon average SPF inflation predictions inform estimates of trend inflation; (ii) inflation gap persistence is countercyclical before the Volcker disinflation and acyclical afterwards; (iii) by 1990 sticky‐information inflation forecast updating is less frequent than it was earlier in the sample; and (iv) the drop in the frequency of the sticky‐information forecast updating occurs at the same time persistent shocks become less important for explaining movements in inflation. Our findings support the view that stickiness in survey forecasts is not invariant to the inflation process.

Estimating the Output Gap for Emerging Countries: Evidence from Five Southeast Asia Countries

This paper presents an application of a bi-variate unobserved components model of output and inflation to estimate the output gap for five emerging economies in Southeast Asia, including the Philippines, Malaysia, Vietnam, Indonesia, and Thailand. In this paper, time-varying stochastic volatility terms are added in the model to exhibit the change in the size of shocks to the trend and cyclical components of output and inflation of these economies. The results show that estimated output gaps are able to identify the recessions of these economies. Although the shape and magnitude of estimated gaps differ from country to country, these gaps imply that these economies tend to converge to a more stable business cycle over time, except for that of Malaysia. Secondly, inflation is very sensitive to the output gap in Vietnam and Indonesia, but not to those in the Philippines, Malaysia, and Thailand. Thirdly, results suggest that time-varying stochastic volatility is clearly seen in the innovations to the output gaps of the five emerging economies, and Indonesia's potential output. Meanwhile, results confirm that there is no need to add stochastic volatility terms in trend components of output and inflation, except for Indonesia's potential output.

Resource abundance, financial crisis and economic growth: did resource‐rich countries fare better during the global financial crisis?

This study examines the role of resource abundance in the cross‐country differences in the impacts of the global financial crisis (GFC) of 2008–2009. Using forecasts from the unobserved components model and exponential smoothing technique, we estimate the output levels a country would reach in 2009 and 2013 in the absence of the GFC, and compare these with the realised output levels. We find large variations in the output losses across 72 countries. The mineral‐rich countries have been found to be in a strong position to survive any adverse shocks stemming from the GFC. Income per capita, trade openness, and institutional quality and government effectiveness are also found to be key factors determining the differences in output loss in the post‐crisis period. These findings have strong implications for resource‐rich countries such as Australia and are expected to shed new light on alternative policy designs and appropriate strategies to deal with any future economic crisis.

A Study on the Development of Global Competitiveness Index for Local Governments in Korea

Purpose - Since 2008, A.T. Kearney has published a Global Cities Index (GCI), which has been evaluated in five areas and 27 detailed evaluation indicators, in 60 major cities in 40 countries. In this study, we selected various indicators necessary for the construction of global competitiveness index of local governments in Korea based on the evaluation system of A.T. Kearney. Through the process of indicator selection, normalization, and aggregation using various weights the global competitiveness index of 16 local governments were calculated, compared and analyzed. Research, design, data and methodology - The study selected 13 indicators of 16 local governments to construct the global competitiveness index. In normalization of transforming individual indicators the parametric method of z-score standardization and nonparametric method which is free from the distribution types are used. Nonparametric methods used in this study include rescale with respect to the maximum value, range standardization, and inter-decile range standardization. In aggregation the global competitiveness index was calculated using weights derived from Analytic Hierarchy Process(AHP), Principal Component Analysis(PCA) and Unobserved Components Model(UCM). Results - The rankings of global competitiveness index using various combinations of normalization method and weight-determination method were calculated and compared. The correlation between the ranks was statistically significant, which implies that all measurement methods are meaningful and useful. It was also found that the method of standardization does not affect the ranking of the composite index while the method of determining the weights affects the ranking of the composite index. Although both the PCA and UCM methods have their advantages and disadvantages in determining weights and making composite index, it is judged that individual indicators can be selected by PCA and then the composite index are developed by UCM in terms of the usefulness of the study. Conclusions - The method for aggregating indicators used in this study is designed to calculate the level, ranking, and confidence interval of a global competitive index. Therefore, it enables us to overcome the limitations of simply calculating the level an ranking of a composite index by arbitrary weight. This method is expected to be very useful in other areas of study.

Is the slope of the Phillips curve time-varying? Evidence from unobserved components models

This paper formally tests for time variation in the slope of the Phillips curve using a variety of measures of inflation expectations and real economic slack. We find that time variation in the slope of the Phillips curve depends on the measure of inflation expectations rather than the measure of real economic slack. We find strong evidence in support of the time-varying slopes of the Phillips curve with different measures of inflation expectations. Thus, we conclude that the slope of the Phillips curve is time-varying.

Unobserved Component Model for Predicting Monthly Traffic Volume

AbstractTraffic volume prediction plays a critical role in transportation system and infrastructure management. This paper develops the first application of an unobserved component model (UCM) for ...

GDP-employment decoupling in Germany

This paper investigates the time-varying relationship between German output and employment growth, in particular their decoupling in recent years. We estimate a correlated unobserved components model that allows for persistent and cyclical time variation in the employment-GDP linkage as well as an additional employment component beyond the one linked to GDP. Controlling for the latter yields a more precise classification of what is a jobless recovery or a labour hoarding recession. We find that the comovement of employment and GDP has loosened permanently since the Great Recession while the comovement with other variables than GDP has become tighter. The decoupling goes hand in hand with the change of the sectoral composition of the economy, especially the rise of the service sector. Beyond that, high labour market tightness and at the same time high labour supply carried recent employment growth. The industry mix, wages and working time had a smaller influence.

Unobserved Components Model for Forecasting Sugarcane Yield in Haryana

Unlike classical regression analysis, the state space models have time-dependent parameters and provide a flexible class of dynamic and structural time series models. The unobserved component model (UCM) is a special type of state space models widely used to analyze and forecast time series. The present investigation has been carried out to study the trend of sugarcane(gur) yield in five districts (Ambala, Karnal, Panipat, Yamunanagar and Kurukshetra) of Haryana state using the unobserved component models with level, trend and irregular components. For this purpose, the time series data on sugarcane yield from 1966-67 to 2016-17 of Ambala and Karnal, 1971-72 to 2016-17 of Kurukshetra and 1980-81 to 2016-17 of Panipat and Yamunanagar districts have been used. For all the districts, the irregular component was found to be highly significant (p=0.01) while both level and trend component variances were observed non-significant. Significance analysis of the individual component(s) has also been performed for possible dropping of the level and trend components by setting their variances equal to zero. The state space models may be effectively used pertaining to Indian agriculture data, as it takes into account the time dependency of the underlying parameters which may further enhance the predictive accuracy of the most popularly used ARIMA models with parameter constancy. Moreover, the unobserved component model is capable of handling both stationary as well as non-stationary time series and thus found more suitable for sugarcane yield modeling which is a trended yield (i.e. non-stationary in nature).