Search

Forecast reconciliation is a post-forecasting process aimed to improve the quality of the base forecasts for a system of hierarchical/grouped time series (Hyndman et al., 2011). Contemporaneous (cross-sectional) and temporal hierarchies have been considered in the literature, but - except for Kourentzes and Athanasopoulos (2019) - generally these two features have not been fully considered together. Adopting a notation able to simultaneously deal with both forecast reconciliation dimensions, the paper shows two new results: (i) an iterative cross-temporal forecast reconciliation procedure which extends, and overcomes some weaknesses of, the two-step procedure by Kourentzes and Athanasopoulos (2019), and (ii) the closed-form expression of the optimal (in least squares sense) point forecasts which fulfill both contemporaneous and temporal constraints. The feasibility of the proposed procedures, along with first evaluations of their performance as compared to the most performing `single dimension' (either cross-sectional or temporal) forecast reconciliation procedures, is studied through a forecasting experiment on the 95 quarterly time series of the Australian GDP from Income and Expenditure sides considered by Athanasopoulos et al. (2019)., Main text: 49 pages, 10 figures, 2 tables. Appendix: 68 pages, 29 figures, 17 tables

Thus far, the focus in prediction market research has been on establishing its forecast accuracy relative to those of other prediction methods, or on the investigation of a few single sources of forecast error. This article is the first attempt to overcome the narrow focus of the literature by combining observational and experimental analyses of prediction market errors. It investigates the prediction error of a real money prediction market uusing a logarithmic market scoring rule for 65 direct democratic votes in Switzerland. The article distinguishes between prediction market error due to the setup of the market, features of the event to be predicted, and the participants involved, and finds that the prediction market accuracy varies primarily according to the setup of the market, with the features of the event and especially the composition of the participant sample hardly mattering.

We provide an extensive study assessing whether a prior wavelet-based denoising step enhances the forecast accuracy of standard forecasting models. Many combinations of attribute values of the thresholding (denoising) algorithm are explored together with several traditional forecasting models used in economic time series forecasting. The results are evaluated using M3 competition yearly time series. We conclude that the performance of a forecasting model combined with the prior denoising step is generally not recommended, which implies that a straightforward generalisation of some of the results available in the literature (which found the denoising step to be beneficial) is not possible. Even if cross-validation is used to select the value of the threshold, a superior performance of the forecasting model with the prior denoising step does not generally follow.

IAS 31 allowed firms to choose between proportionate consolidation and the equity method to record joint ventures in the consolidated accounts of the venturer. Moreover, this election implied a decision about including information in the primary financial statements or in the notes. This paper investigates if financial analysts perceive accounting information differently depending on the method chosen conditioned to the disclosure of the required information in the notes. We analysed a sample of Spanish firms during 2005–2010. We not only considered earnings forecasts, but also examined target prices and stock recommendations. Furthemore, we look at how this accounting choice affects analysts’ information environment. Our results suggest that the choice of accounting regime does not affect the bias and accuracy of earnings forecasts, nor target prices nor stock recommendations, no matter if firms provide or not information in the notes. While the proportionate method implies lower dispersion in analysts’ forecasts than the equity method, our tests do not allow us to confirm that the information environment depends on the accounting method. These results support the decision adopted in IFRS 11 to impose a unique method for the accounting of joint ventures. La NIC 31 permitía elegir entre integración proporcional y el método de la participación para la inclusión de los negocios conjuntos en las cuentas consolidadas de los partícipes. Además, esta elección conllevaba la inclusión de cierta información sobre los negocios conjuntos en los estados financieros primarios o en la memoria. Este trabajo investiga si los analistas financieros perciben la información contable de forma distinta dependiendo del método elegido, condicionado a que se divulgue la información requerida en la memoria. Se analiza una muestra de empresas españolas en el periodo 2005-2010, y se considera la influencia de la decisión contable sobre las predicciones de beneficios, precios objetivo y recomendaciones sobre acciones. Adicionalmente, se estudia si esta opción contable afecta al entorno informativo de los analistas. Nuestros resultados sugieren que la elección no afecta al sesgo ni a la precisión de las predicciones de beneficios, ni tampoco a precios objetivo ni a recomendaciones, al margen de que las empresas proporcionen o no información en la memoria. Pese a que la integración proporcional implica menor dispersión en las predicciones, los contrastes realizados no permiten confirmar que el entorno informativo se vea afectado por el uso de un método contable u otro. Estos resultados apoyan la decisión tomada en la NIIF 11 de imponer un único método para la contabilización de los negocios conjuntos. Begoña Giner gratefully acknowledges the financial support of the Spanish Ministry of Economy and Competitiveness, under grant [ECO2013-48208-P].

Portfolio theory assumes that investors accept risk. This means thatin the equal rate of return on the two assets, the assets were chosenthat have a lower risk level. Modern portfolio theory is accepted byinvestors who believe that they are not cope with the market. Sothey keep many different types of securities in order to access theoptimum efficiency rate that is close to the rate of return on market.One way to control investment risk is establishing the portfolioshares. There are many ways to choose the optimal portfolioshares. Among these methods in this study we use loss functions.For this, we choose all firms from the year2011to the end of 2015that had been a member in the Tehran Stock Exchange. The resultsof this research show that the likelihood functions have the bestperformance in Forecasting the optimal portfolio allocationprob-lem.

Post-2006 expert judgment data has been extended to 530 experts assessing 580 calibration variables from their fields. New analysis shows that point predictions as medians of combined expert distributions outperform combined medians, and medians of performance weighted combinations outperform medians of equal weighted combinations. Relative to the equal weight combination of medians, using the medians of performance weighted combinations yields a 65% improvement. Using the medians of equally weighted combinations yields a 46% improvement. The Random Expert Hypothesis underlying all performance-blind combination schemes, namely that differences in expert performance reflect random stressors and not persistent properties of the experts, is tested by randomly scrambling expert panels. Generating distributions for a full set of performance metrics, the hypotheses that the original panels’ performance measures are drawn from distributions produced by random scrambling are rejected at significance levels ranging from E − 6 to E − 12 . Random stressors cannot produce the variations in performance seen in the original panels. In- and out-of-sample validation results are updated.

The generalized smooth transition autoregression (GSTAR) parametrizes the joint asymmetry in the duration and length of cycles in macroeconomic time series by using particular generalizations of the logistic function. The symmetric smooth transition and linear autoregressions are nested in the GSTAR. A test for the null hypothesis of dynamic symmetry is presented. Two case studies indicate that dynamic asymmetry is a key feature of the U.S. economy. The GSTAR model beats its competitors for point forecasting, but this superiority becomes less evident for density forecasting and in uncertain forecasting environments.

We investigate how the forecasting performance of the Federal Reserve Greenbooks has changed relative to commercial forecasters between 1974 and 2009. To this end, we analyze time-variation in the Greenbook coefficients in forecast encompassing regressions. Assuming that model coefficients change continuously, we estimate unobserved components models using Bayesian inference techniques. To verify that our results do not depend on the specific way change is modeled, we also allow the coefficients to change discretely rather than continuously and test for structural breaks using classical inference techniques. We find that the Greenbook forecasts have been consistently superior to the commercial forecasts at all horizons throughout our sample period. Although the forecasting performance gap has narrowed at more distant horizons after the early-to-mid 1980s, it remains significant.

Using stochastic forecasting techniques, this paper assesses the consequences for public finances of changes in age and household structures in Denmark over the period 2008–2037. Focusing on components of welfare provisions and tax payments with noticeable differences across age and household status, we show that, based on a point forecast, the fiscal impact of changes in household structures amounts to an annual negative effect of 0.5% of GDP, and the effect of changes in age structures is forecast to worsen the public budget by 3.7% of GDP per year. While being subject to a considerable amount of uncertainty, the prospect of such a dramatic weakening of public finances is likely to trigger demands for welfare reforms characterized by a more individualized system of public transfer and tax payments, in addition to the measures that have already been taken to address the fiscal effects of population ageing.

This study uses quantile regressions to estimate historical forecast error distributions for WASDE forecasts of corn, soybean, and wheat prices, and then compute confidence limits for the forecasts based on the empirical distributions. Quantile regressions with fit errors expressed as a function of forecast lead time are consistent with theoretical forecast variance expressions while avoiding assumptions of normality and optimality. Based on out-of-sample accuracy tests over 1995/96–2006/07, quantile regression methods produced intervals consistent with the target confidence level. Overall, this study demonstrates that empirical approaches may be used to construct accurate confidence intervals for WASDE corn, soybean, and wheat price forecasts.

Probability forecasts of the Swiss franc/euro (CHF/EUR) exchange rate are generated before, surrounding and after the placement of a floor on the CHF/EUR by the Swiss National Bank (SNB). The goal is to determine whether the exchange rate floor has a positive, negative or insignificant effect on the calibration of the probability forecasts from three time-series models: a vector autoregression (VAR) model, a VAR model augmented with the LiNGAM causal learning algorithm, and a univariate autoregressive model built on the independent components (ICs) of an independent component analysis (ICA). Score metric rankings of forecasts and plots of calibration functions are used in an attempt to identify the preferred time-series model based on forecast performance. The study not only finds evidence that the floor on the CHF/EUR has a negative impact on the forecasting performance of all three time-series models but also that the policy change by the SNB altered the causal structure underlying the six major currencies.

The paper compares the information contents of two of the most widely used sets of macroeconomic forecasts, namely, the macroeconomic forecasts produced by the International Monetary Fund (IMF) and the Organization for Economic Cooperation and Development (OECD). By testing whether the release of public sector forecasts had systematic impact on subsequent revisions in Consensus (private sector) forecasts, the paper finds that, for the period 1994?2003, private forecasters reacted to the information provided by the IMF on regions (such as non?G7 Europe and Latin America) for which the IMF’s forecasts are known to be more accurate or about which information is less available. What determines the information value of public sector forecasts, however, appears complex. The OECD’s impact was limited in influencing private sector forecasts for the G7 countries, though it had a superior forecast performance than the IMF. Some herding behavior was observed among private sector forecasters.

It is common practice to evaluate fixed-event forecast revisions in macroeconomics by regressing current forecast revisions on one-period lagged forecast revisions. Under weak-form (forecast) efficiency, the correlation between the current and one-period lagged revisions should be zero. The empirical findings in the literature suggest that this null hypothesis of zero correlation is rejected frequently, where the correlation can be either positive (which is widely interpreted in the literature as “smoothing”) or negative (which is widely interpreted as “over-reacting”). We propose a methodology to interpret such non-zero correlations in a straightforward and clear manner. Our approach is based on the assumption that numerical forecasts can be decomposed into both an econometric model and random expert intuition. We show that the interpretation of the sign of the correlation between the current and one-period lagged revisions depends on the process governing intuition, and the current and lagged correlations between intuition and news (or shocks to the numerical forecasts). It follows that the estimated non-zero correlation cannot be given a direct interpretation in terms of smoothing or overreaction.

256 σ., Αντικείμενο της παρούσας διδακτορικής διατριβής είναι η ανάλυση των παραγόντων που επηρεάζουν τη στατιστική ακρίβεια πρόβλεψης των τεχνικών προέκτασης. Η μελέτη εστιάζει στα δεδομένα συνεχούς φύσης, τα οποία ορίζονται ως μία συνάρτηση των τεσσάρων βασικών συνιστωσών τους, που είναι η εποχιακότητα, η τάση, ο κύκλος και η τυχαιότητα. Η βιβλιογραφική έρευνα που εκπονήθηκε στα πλαίσια της διατριβής είχε ως στόχο την πλήρη κατανόηση των χρονοσειρών καθώς και της ανάλυσης αυτών. Μελετήθηκαν τα κύρια ποιοτικά χαρακτηριστικά, και έγινε καταγραφή μεθόδων αποσύνθεσης, στατιστικής ανάλυσης και μετασχηματισμού των αρχικών δεδομένων. Κατόπιν, μελετήθηκε σε βάθος η βιβλιογραφία που αφορά καθιερωμένες και σύγχρονες μεθοδολογίες πρόβλεψης, τόσο σε επίπεδο χρονοσειρών συνεχούς ζήτησης, όσο και σε επίπεδο δεδομένων διακοπτόμενης φύσης. Μελετήθηκε και αναπτύχθηκε πλήρης βιβλιογραφία σχετικά με τους παράγοντες που επηρεάζουν τις προβλέψεις, διακρίνοντάς τρεις βασικές κατηγορίες: στατιστικούς, κριτικούς και ψυχολογικούς. Η έρευνα επί του μετασχηματισμού οδήγησε στην περιγραφή και παρουσίαση της καινοτόμας μεθόδου προβλέψεων ADIDA, που πρόκειται για μια τεχνική συνάθροισης των δεδομένων σε σειρές χαμηλότερης συχνότητας με στόχο τη μείωση της παρουσίας μηδενικών τιμών. Η συνάθροιση των δεδομένων ακολουθείται από πρόβλεψη στο συναθροισμένο επίπεδο και, τελικά, διαχωρισμό των συναθροισμένων προβλέψεων σε προβλέψεις υψηλότερης συχνότητας. Η φιλοσοφία αυτή εφαρμόστηκε σε σύνολα πραγματικών δεδομένων συνεχούς και διακοπτόμενης φύσης και επιδεικνύει πολύ ελπιδοφόρα αποτελέσματα, λειτουργώντας ως ένας μηχανισμός «αυτό-βελτίωσης» των τεχνικών προέκτασης. Η μελέτη, ανάλυση και εξέταση των παραγόντων που επηρεάζουν τις στατιστικές προβλέψεις επιτεύχθηκε μέσω μιας εκτενούς πειραματικής διαδικασίας προσομοίωσης. Η διατριβή περιγράφει αναλυτικά την επιλογή των συνιστωσών και των επιπέδων αυτών αλλά και τη διαδικασία κατασκευής ενός πολύ μεγάλου τεχνητού συνόλου δεδομένων (σχεδόν 80 εκατομμύρια σειρές), οι οποίες και αποτέλεσαν και τη βάση για το κύριο μέρος της έρευνας της παρούσας διατριβής. Η στατιστική ανάλυση μέσω πολλαπλής παλινδρόμησης καθιστά σαφές το ύψος της επιρροής των εξεταζόμενων παραγόντων. Τέλος, προτείνεται ένα πλαίσιο επιλογής μεθόδου βάσει της ανάλυσης των χαρακτηριστικών της χρονοσειράς., The current doctoral thesis’ main objective is the analysis of factors affecting statistical accuracy of extrapolation techniques. The study focuses on fast demand time series, which can be described as a function of four basic components: seasonality, trend, cycle and randomness. The literature review conducted in this thesis was to understand every aspect of time series analysis. We studied the main qualitative characteristics and we recorded widely used methods regarding decomposition, statistical analysis and transformation of the original data. Furthermore, we studied in depth the literature on well-established and modern forecasting methodologies, regarding on both fast and intermittent demand natures. Lastly, we studied and developed comprehensive literature review on factors affecting forecasts and forecasting procedure, by distinguishing three main categories: statistical, judgmental and psychological factors. The research regarding transformation of the original data led to the definition of a unique forecasting methodology, the ADIDA framework, which is an aggregation-disaggregation technique of gathering data in lower frequency ranges, so to reduce dramatically the presence of zero values. The aggregation of data is followed by extrapolation at the aggregated level and, finally, separation of the aggregated point forecast at forecasts of higher frequency. This philosophy was applied to real data sets of fast as well as intermittent demand nature, where the results were very promising, serving as a “self-improvement” mechanism for forecasting methods. The analysis and examination of factors affecting the statistical predictions was achieved through an extensive experimental simulation process. The thesis describes the selection of the levels for each examined factor and the generation procedure of an extremely large simulated data set (including about 80 million series), which was the basis for the main part of the current research. The statistical analysis by multiple regression functions verifies the main hypothesis and illustrates the level of influence of the examined factors. Finally, we propose a method selection framework, based on the analysis of series characteristics and the results of this study.

In this paper we face the fitting versus forecasting paradox with the objective of realizing an optimal Early Warning System to better describe and predict past and future sovereign defaults. We do this by proposing a new Regression Tree-based model that signals a potential crisis whenever preselected indicators exceed specific thresholds. Using data on 66 emerging markets over the period 1975-2002, our model provides an accurate description of past data, although not the best description relative to existing competing models (Logit, Stepwise logit, Noise-to-Signal Ratio and Regression Trees), and produces the best forecasts accomodating to different risk aversion targets. By modulating in- and out-of sample model accuracy, our methodology leads to unambiguous empirical results, since we find that illiquidity (short-term debt to reserves ratio), insolvency (reserve growth) and contagion risks act as the main determinants/predictors of past/future debt crises.

This paper investigates the forecasting ability of survey data on exchange rate expectations with multiple forecast horizons. The survey forecasts are on the exchange rates of five Central and Eastern European currencies: Czech Koruna, Hungarian Forint, Polish Zloty, Romanian Leu and Slovakian Koruna. First, different term-structure models are fitted on the survey forecasts. Then, the forecasting performances of the fitted forecasts are compared. The fitted forecasts for the 5 months horizon and beyond are proved to be significantly better than the random walk on the pooled data of the five currencies. The best performing term-structure model is the one that assumes an exponential relationship between the forecast and the forecast horizon, and has time-varying parameters.

Nowcasting has been a challenge in the recent economic crisis. We introduce the Toll Index, a new monthly indicator for business cycle forecasting and demonstrate its relevance using German data. The index measures the monthly transportation activity performed by heavy transport vehicles across the country and has highly desirable availability properties (insignificant revisions, short publication lags) as a result of the innovative technology underlying its data collection. It is coincident with production activity due to the prevalence of just-in-time delivery. The Toll Index is a good early indicator of production as measured for instance by the German Production Index, provided by the German Statistical Office, which is a well-known leading indicator of the Gross National Product. The proposed new index is an excellent example of technological, innovation-driven economic telemetry, which we suggest should be established more around the world.

It is common practice to evaluate fixed-event forecast revisions in macroeconomics by regressing current revisions on one-period lagged revisions. Under weak-form efficiency, the correlation between the current and one-period lagged revisions should be zero. The empirical findings in the literature suggest that the null hypothesis of zero correlation between the current and one-period lagged revisions is rejected quite frequently, where the correlation can be either positive or negative. In this paper we propose a methodology to be able to interpret such non-zero correlations in a straightforward manner. Our approach is based on the assumption that forecasts can be decomposed into both an econometric model and expert intuition. The interpretation of the sign of the correlation between the current and one-period lagged revisions depends on the process governing intuition, and the correlation between intuition and news.

When forecasting decisions in conflict situations, experts are often advised to figuratively stand in the other person’s shoes. We refer to this as “role thinking”, because, in practice, the advice is to think about how other protagonists will view the situation in order to predict their decisions. We tested the effect of role thinking on forecast accuracy. We obtained 101 role-thinking forecasts of the decisions that would be made in nine diverse conflicts from 27 Naval postgraduate students (experts) and 107 role-thinking forecasts from 103 second-year organizational behavior students (novices). The accuracy of the novices’ forecasts was 33% and that of the experts’ was 31%; both were little different from chance (guessing), which was 28%. The small improvement in accuracy from role-thinking strengthens the finding from earlier research that it is not sufficient to think hard about a situation in order to predict the decisions which groups of people will make when they are in conflict. Instead, it is useful to ask groups of role players to simulate the situation. When groups of novice participants adopted the roles of protagonists in the aforementioned nine conflicts and interacted with each other, their group decisions predicted the actual decisions with an accuracy of 60%. Refereed/Peer-reviewed

Models for the twelve-month-ahead US rate of inflation, measured by the chain weighted consumer expenditure deflator, are estimated for 1974-99 and subsequent pseudo out-of-sample forecasting performance is examined. Alternative forecasting approaches for different information sets are compared with benchmark univariate autoregressive models, and substantial out-performance is demonstrated. Three key ingredients to the out-performance are: including equilibrium correction terms in relative prices; introducing non-linearities to proxy state dependence in the inflation process; and replacing the information criterion, commonly used in VARs to select lag length, with a ‘parsimonious longer lags’ (PLL) parameterisation. Forecast pooling or averaging also improves forecast performance.

Inflation is a far from homogeneous phenomenon, a fact often neglected in modelling consumer price inflation. This study, the first of its kind for an emerging market country, investigates gains to inflation forecast accuracy by aggregating weighted forecasts of the sub-component price indices, versus forecasting the aggregate consumer price index itself. Rich multivariate equilibrium correction models employ general and sectoral information for ten sub-components, taking account of structural breaks and institutional changes. Model selection is over 1979-2003, with pseudo out-of-sample forecasts, four-quarters-ahead, generated to 2007. Aggregating the weighted forecasts of the sub-components does outperform the aggregate CPI forecasts, and also offers substantial gains over forecasting using benchmark naïve models. The analysis also contributes an improved understanding of sectoral inflationary pressures. This forecasting method should be more robust to the regular reweighting of the CPI index.

We analyze forecasts of inflation and GDP growth contained in Banco de México's Survey of Professional Forecasters for the period 1995-2009. The forecasts are for the current and the following year, comprising an unbalanced three-dimensional panel with multiple individual forecasters, target years, and forecast horizons. The fixed-event nature of the forecasts enables us to examine efficiency by looking at the revision process. The panel structure allows us to control for aggregate shocks and to construct a measure of the news that impacted expectations in the period under study. The results suggest that respondents seem to rely for longer than appears to be optimal on their previous forecasts, and that they do not seem to use past information in an efficient manner. In turn, this means there are areas of opportunity to improve the accuracy of the forecasts, for instance, by taking into account the positive autocorrelation found in forecast revisions.

This paper proposes a strategy to increase the efficiency of forecast combination. Given the availability of a wide range of forecasts for the same variable of interest, our goal is to apply combining methods to a restricted set of models. With this aim, a hierarchical procedure based on an encompassing test is considered. First, forecasting models are ranked according to a measure of predictive accuracy (RMSFE). The models are then selected for combination such that each forecast is not encompassed by any of the competing forecasts. Thus the hierarchical procedure represents a compromise between model selection and model averaging. The robustness of the procedure is investigated in terms of the relative RMSFE using ISAE (Institute for Studies and Economic Analyses) short-term forecasting models for monthly industrial production in Italy.

In this paper, we use survey data to analyze the accuracy, unbiasedness, and the efficiency of professional macroeconomic forecasts. We analyze a large panel of individual forecasts that has not been analyzed in the literature so far. We provide evidence on the properties of forecasts for all G7 counties and for four diffierent macroeconomic variables. Our results show a high degree of dispersion of forecast accuracy across forecasters. We also find that there are large diffierences in the performance of forecasters not only across countries but also across diffierent macroeconomic variables. In general, forecasts tend to be biased in situations where forecasters have to respond to large structural shocks or gradual changes in the trend of a variable. Furthermore, while a sizable fraction of forecasters seem to smooth their GDP forecasts significantly, this does not apply to forecasts made for other macroeconomic variables.

This paper proposes a strategy to increase the efficiency of forecast combination. Given the availability of a wide range of forecasts for the same variable of interest, our goal is to apply combining methods to a restricted set of models. To this aim, a hierarchical procedure based on an encompassing test is developed. Firstly, forecasting models are ranked according to a measure of predictive accuracy (RMSFE). The models are then selected for combination such that each forecast is not encompassed by any of the competing forecasts. Thus, the procedure aims to unit model selection and model averaging methods. The robustness of the procedure is investigated in terms of the relative RMSFE using ISAE (Institute for Studies and Economic Analyses) short-term forecasting models for monthly industrial production in Italy.