nep-for New Economics Papers
on Forecasting
Issue of 2007‒12‒19
nine papers chosen by
Rob J Hyndman
Monash University

  1. Bayesian Forecast Combination for VAR Models By Andersson, Michael K; Karlsson, Sune
  2. The Riksbank’s Forecasting Performance By Andersson, Michael K.; Karlsson, Gustav; Svensson, Josef
  3. Polar Bear Population Forecasts: A Public-Policy Forecasting Audit By Armstrong, J. Scott; Green, Kesten C.; Soon, Willie
  4. Does global liquidity help to forecast US inflation? By D'Agostino, A; Surico, P
  5. Joint Modeling of Call and Put Implied Volatility By Ahoniemi, Katja; Lanne, Markku
  6. Which cost of debt should be used in forecasting cash flows? By Ignacio Velez-Pareja
  7. Prospective analysis: guidelines for forecasting financial statements By Ignacio Velez-Pareja; Joseph Tham
  8. To plug or not to plug, that is the question. No plugs, no circularity: A better way to forecast financial statements By Ignacio Velez-Pareja
  9. Predictions by early indicators of the time and height of yearly influenza outbreaks in Sweden By Andersson, Eva; Kühlmann-Berenzon, Sharon; Linde, Annika; Schiöler, Linus; Rubinova, Sandra; Frisén, Marianne

  1. By: Andersson, Michael K (Department of Business, Economics, Statistics and Informatics); Karlsson, Sune (Department of Business, Economics, Statistics and Informatics)
    Abstract: We consider forecast combination and, indirectly, model selection for VAR models when there is uncertainty about which variables to include in the model in addition to the forecast variables. The key dierence from traditional Bayesian variable selection is that we also allow for uncertainty regarding which endogenous variables to include in the model. That is, all models include the forecast variables, but may otherwise have diering sets of endogenous variables. This is a dicult problem to tackle with a traditional Bayesian approach. Our solution is to focus on the forecasting performance for the variables of interest and we construct model weights from the predictive likelihood of the forecast variables. The procedure is evaluated in a small simulation study and found to perform competitively in applications to real world data.
    Keywords: Bayesian model averaging; Predictive likelihood; GDP forecasts
    JEL: C11 C15 C32 C52 C53
    Date: 2007–12–13
  2. By: Andersson, Michael K. (Monetary Policy Department, Central Bank of Sweden); Karlsson, Gustav (Monetary Policy Department, Central Bank of Sweden); Svensson, Josef (Monetary Policy Department, Central Bank of Sweden)
    Abstract: This paper describes the official Riksbank forecasts for the period 2000-06. The forecast variables are those that are important for monetary policy analysis, i.e. inflation, GDP, productivity, employment, labour force, unemployment and financial variables such as interest rate and foreign exchange rate. The Riksbank’s forecasts are presented and analyzed and compared with alternative forecasts, that is, those from other institutions and simple statistical models. One important message from the study is that macroeconomic forecasts are associated with an appreciable uncertainty; the forecast errors are often sizeable. The forecast memory, defined as how far the forecasts are more informative than the variables unconditional mean, is usually limited to the first year. Furthermore, we find that the inflation forecasts exhibit several appealing features, such as a predictability memory that (possibly) includes the second year, relatively low RMSE and weak efficiency. The forecasts for the investigated real variables are shown to be less precise and they have a shorter forecast memory. The exchange rate predictions demonstrate the least accurate (of the investigated variables) forecasts. Compared to other forecasters, the Riksbank’s predictions are often more accurate. This holds for a comparison with the National Institute of Economic Research, even though the differences are statistically insignificant, as well as for a comparison with the participants in the Consensus Forecasts panel, where the Riksbank’s predictions often are among the best. We also find indications that misjudgements for productivity growth have had effects on forecasts for both inflation and GDP, but the results suggest that the Riksbank has considered available information in an acceptable fashion. This is also true for the undertaken revisions (from one forecast occasion to another) of the published forecasts.
    Keywords: Judgements; Forecast Evaluation; Central Bank; Inflation; GDP; RMSE
    JEL: E27 E37 E52
    Date: 2007–12–01
  3. By: Armstrong, J. Scott; Green, Kesten C.; Soon, Willie
    Abstract: The extinction of polar bears by the end of the 21st century has been predicted and calls have been made to list them as a threatened species under the U.S. Endangered Species Act. The decision on whether or not to list rests upon forecasts of what will happen to the bears over the 21st Century. Scientific research on forecasting, conducted since the 1930s, has led to an extensive set of principles—evidence-based procedures—that describe which methods are appropriate under given conditions. The principles of forecasting have been published and are easily available. We assessed polar bear population forecasts in light of these scientific principles. Much research has been published on forecasting polar bear populations. Using an Internet search, we located roughly 1,000 such papers. None of them made reference to the scientific literature on forecasting. We examined references in the nine unpublished government reports that were prepared “…to Support U.S. Fish and Wildlife Service Polar Bear Listing Decision.” The papers did not include references to works on scientific forecasting methodology. Of the nine papers written to support the listing, we judged two to be the most relevant to the decision: Amstrup, Marcot and Douglas et al. (2007), which we refer to as AMD, and Hunter et al. (2007), which we refer to as H6 to represent the six authors. AMD’s forecasts were the product of a complex causal chain. For the first link in the chain, AMD assumed that General Circulation Models (GCMs) are valid. However, the GCM models are not valid as a forecasting method and are not reliable for forecasting at a regional level as being considered by AMD and H6, thus breaking the chain. Nevertheless, we audited their conditional forecasts of what would happen to the polar bear population assuming that the extent of summer sea ice will decrease substantially in the coming decades. AMD could not be rated against 26 relevant principles because the paper did not contain enough information. In all, AMD violated 73 of the 90 forecasting principles we were able to rate. They used two un-validated methods and relied on only one polar bear expert to specify variables, relationships, and inputs into their models. The expert then adjusted the models until the outputs conformed to his expectations. In effect, the forecasts were the opinions of a single expert unaided by forecasting principles. Based on research to date, approaches based on unaided expert opinion are inappropriate to forecasting in situations with high complexity and much uncertainty. Our audit of the second most relevant paper, H6, found that it was also based on faulty forecasting methodology. For example, it extrapolated nearly 100 years into the future on the basis of only five years of data – and data for these years were of doubtful validity. In summary, experts’ predictions, unaided by evidence-based forecasting procedures, should play no role in this decision. Without scientific forecasts of a substantial decline of the polar bear population and of net benefits from feasible policies arising from listing polar bears, a decision to list polar bears as threatened or endangered would be irresponsible.
    Keywords: adaptation; bias; climate change; decision making; endangered species; expert opinion; evaluation; evidence-based principles; expert judgment; extinction; forecasting methods; global warming; habitat loss; mathematical models; scientific method; sea ice
    JEL: C53 H0 C5 C0 H23 C4
    Date: 2007–12–15
  4. By: D'Agostino, A; Surico, P
    Abstract: We construct a measure of global liquidity using the growth rates of broad money for the G7 economies. Global liquidity produces forecasts of US inflation that are significantly more accurate than the forecasts based on US money growth, Phillips curve, autoregressive and moving average models. The marginal predictive power of global liquidity is strong at three years horizons. Results are robust to alternative measures of inflation.
    JEL: C53 C22 E37 E47
    Date: 2007–11
  5. By: Ahoniemi, Katja; Lanne, Markku
    Abstract: This paper exploits the fact that implied volatilities calculated from identical call and put options have often been empirically found to differ, although they should be equal in theory. We propose a new bivariate mixture multiplicative error model and show that it is a good fit to Nikkei 225 index call and put option implied volatility (IV). A good model fit requires two mixture components in the model, allowing for different mean equations and error distributions for calmer and more volatile days. Forecast evaluation indicates that in addition to jointly modeling the time series of call and put IV, cross effects should be added to the model: putside implied volatility helps forecast callside IV, and vice versa. Impulse response functions show that the IV derived from put options recovers faster from shocks, and the effect of shocks lasts for up to six weeks.
    Keywords: Implied Volatility; Option Markets; Multiplicative Error Models; Forecasting
    JEL: C32 C53 G13
    Date: 2007
  6. By: Ignacio Velez-Pareja
    Abstract: Frequently analysts and teachers use the capitalized rate of interest for the cost of debt when forecasting and discounting cash flows. On the other hand, some authors (and analysts) estimate the interest payments when forecasting annual financial statements or cash flows based on the average of debt calculated with the beginning balance and the end of year balance. This makes some sense because usually firms repay debt in a monthly or quarterly basis and calculating interest payments might not reflect reality. Others use the end of year convention that calculates the yearly interest multiplying the beginning balance times the contractual cost of debt. In this teaching note we show the differences when we use those different approaches and make a simple proposal to solve the problem.
    Date: 2007–12–04
  7. By: Ignacio Velez-Pareja; Joseph Tham
    Abstract: We discuss some ideas useful when forecasting financial statements that are based on historical data. The chapter is organized as follows: First we discuss the relevance of prospective analysis for non traded firms. In a second section we a basic reviews of subjects that will be needed for forecasting financial statements. We discuss the use of plugs for financial forecasting. We show an alternate approach to avoid such popular practice. The approach we propose follows the Double Entry Principle. This principle guarantees consistent and error free financial statements. We show with a simple example how the plug works and its limitations and problems that arise when using it. Next, the reader will find what information is needed for the forecasting of financial statements and where and how to find it. We present the procedure to identify policies that govern the ongoing of a firm such as accounts receivable and payable, inventories, dividend payout, and identify price increases and other basic variables. We also deal with the real life problem of a firm with multiple products and/or services. We start with historical financial statements, We include inflation rates, real increases in prices and volume and policies in order to construct intermediate tables that make very easy the construction of the pro forma financial statements. We use a detailed example to illustrate the method. We derive the cash flows that will be used in the book to value a firm. This type of models might be used by non traded firm for a permanent assessment of the value creation. Finally we show some tools to perform sensitivity analysis for financial management and analysis.
    Date: 2007–12–04
  8. By: Ignacio Velez-Pareja
    Abstract: Typical textbooks on corporate finance and forecasting and budgeting recommend “closing” and matching the financial statements using what is known as a plug. A plug is a formula to match the Balance Sheet using differences in some items listed in it in such a way that the accounting equation holds. This is a very easy way to do it but it encompasses some risks. The risks are that certain numbers in the financial statements could be in error and still the plug would indicate that everything is correct because the Balance Sheet matches. In this teaching note we show how to construct financial statement without plugs and circularity.
    Date: 2007–12–04
  9. By: Andersson, Eva (Statistical Research Unit, Department of Economics, School of Business, Economics and Law, Göteborg University); Kühlmann-Berenzon, Sharon (Department of Epidemiology, Swedish Institute for Infectious Disease Control, Stockholm Group for Epidemic Modelling); Linde, Annika (Department of Epidemiology, Swedish Institute for Infectious Disease Control); Schiöler, Linus (Statistical Research Unit, Department of Economics, School of Business, Economics and Law, Göteborg University); Rubinova, Sandra (Department of Epidemiology, Swedish Institute for Infectious Disease Control); Frisén, Marianne (Statistical Research Unit, Department of Economics, School of Business, Economics and Law, Göteborg University)
    Abstract: Aims: Methods for prediction of the peak of the influenza from early observations are suggested. These predictions can be used for planning purposes. Methods: In this study, new robust methods are described and applied on weekly Swedish data on influenza-like illness (ILI) and weekly laboratory diagnoses of influenza (LDI). Both simple and advanced rules for how to predict the time and height of the peak of LDI are suggested. The predictions are made using covariates calculated from data in early LDI reports. The simple rules are based on the observed LDI values while the advanced ones are based on smoothing by unimodal regression. The suggested predictors were evaluated by cross-validation and by application to the observed seasons. Results: The relation between ILI and LDI was investigated and it was found that the ILI variable is not a good proxy for the LDI variable. The advanced prediction rule regarding the time of the peak of LDI had a median error of 0.9 weeks, and the advanced prediction rule for the height of the peak had a median deviation of 28%. Conclusions: The statistical methods for predictions have practical usefulness.
    Keywords: Prediction; Influenza; Outbreak
    JEL: C10
    Date: 2007–01–01

This nep-for issue is ©2007 by Rob J Hyndman. It is provided as is without any express or implied warranty. It may be freely redistributed in whole or in part for any purpose. If distributed in part, please include this notice.
General information on the NEP project can be found at For comments please write to the director of NEP, Marco Novarese at <>. Put “NEP” in the subject, otherwise your mail may be rejected.
NEP’s infrastructure is sponsored by the School of Economics and Finance of Massey University in New Zealand.