Search

We propose a novel copula approach to producing density forecasts of economic aggregates combining models using disaggregate data. Our copula approach is more flexible compared to existing techniques, because it is applicable to any econometric model that produces density forecasts. We construct a set of Monte Carlo studies to investigate the properties of the suggested approach. In our empirical application, we use the Norwegian index for goods consumption (VKI) and the Norwegian consumer price index for underlying inflation (CPI-ATE). We find that the copula approach compares well to alternative methods using recursive out-of-sample estimation.

We use a novel data set covering all domestic debit card transactions in physical terminals by Norwegian households, to nowcast quarterly Norwegian household consumption. These card payments data are free of sampling errors and are available weekly without delays, providing a valuable early indicator of household spending. To account for mixed-frequency data, we estimate various mixed-data sampling (MIDAS) regressions using predictors sampled at monthly and weekly frequency. We evaluate both point and density forecasting performance over the sample 2011Q4-2020Q1. Our results show that MIDAS regressions with debit card transactions data improve both point and density forecast accuracy over competitive standard benchmark models that use alternative high-frequency predictors. Finally, we illustrate the benefits of using the card payments data by obtaining a timely and relatively accurate nowcast of the first quarter of 2020, a quarter characterized by heightened uncertainty due to the COVID-19 pandemic.

Our analysis suggests; they do not! To arrive at this conclusion we construct a real-time data set of interest rate projections from central banks in three small open economies; New Zealand, Norway, and Sweden, and analyze if revisions to these projections (i.e., forward guidance) can be predicted by timely information. Doing so, we find a systematic role for forward looking international indicators in predict- ing the revisions to the interest rate projections in all countries. In contrast, using similar indexes for the domestic economy yields largely insignificant results. Furthermore, we find that revisions to forward guidance matter. Using a VAR identified with external instruments based on forecast errors from the predictive regressions, we show that the responses to output, inflation, the exchange rate and asset returns resemble those one typically associates with a conventional monetary policy shock.

We analyze how to incorporate low frequency information in models for predicting high frequency variables. In doing so, we introduce a new model, the reverse unrestricted MIDAS (RU-MIDAS), which has a periodic structure but can be estimated by simple least squares methods and used to produce forecasts of high frequency variables that also incorporate low frequency information. We compare this model with two versions of the mixed frequency VAR, which so far had been only applied to study the reverse problem, that is, using the high frequency information for predicting low frequency variables. We then implement a simulation study to evaluate the relative forecasting ability of the alternative models in finite samples. Finally, we conduct several empirical applications to assess the relevance of quarterly survey data for forecasting a set of monthly macroeconomic indicators. Overall, it turns out that low frequency information is important, particularly so when it is just released.

We review several methods to define and forecast classical business cycle turning points in Norway. In the paper we compare the Bry - Boschan rule (BB) with a Markov Switching model (MS), using alternative vintages of Norwegian Gross Domestic Product (GDP) as the business cycle indicator. The timing of business cycles depends on the vintage and the method used. BB provides the most reasonable definition of business cycles. The forecasting exercise, where the models are augmented with surveys or financial indicators, respectively, leads to the conclusion that the BB rule applied to density forecasts of GDP augmented with either the consumer confidence index or a financial conditions index provides the most timely predictions of peaks. For troughs, augmenting with surveys or financial indicators does not increase forecastability.

This paper compares alternative models of time-varying macroeconomic volatility on the basis of the accuracy of point and density forecasts of macroeconomic variables. In this analysis, we consider both Bayesian autoregressive and Bayesian vector autoregressive models that incorporate some form of time-varying volatility, precisely stochastic volatility (both with constant and time-varying autoregressive coefficients), stochastic volatility following a stationary AR process, stochastic volatility coupled with fat tails, GARCH and mixture of innovation models. The comparison is based on the accuracy of forecasts of key macroeconomic time series for real-time post War-II data both for the United States and United Kingdom. The results show that the AR and VAR specifications with widely-used stochastic volatility dominate models with alternative volatility specifications, in terms of point forecasting to some degree and density forecasting to a greater degree.

We apply a suite of models to produce quasi-real-time density forecasts of Norwegian GDP and inflation, and evaluate different combination and selection methods using the Kullback-Leibler information criterion (KLIC). We use linear and logarithmic opinion pools in conjunction with various weighting schemes, and we compare these combinations to two different selection methods. In our application, logarithmic opinion pools were better than linear opinion pools, and score-based weights were generally superior to other weighting schemes. Model selection generally yielded poor density forecasts, as evaluated by KLIC.

New-generation DSGE models are sometimes misspecified in dimensions that matter for their forecasting performance. The paper suggests one way to improve the forecasts of a DSGE model using a conditioning information that need not be accurate. The technique presented allows for agents to anticipate the information on the conditioning variables several periods ahead. It also allows the forecaster to apply a continuum of degrees of uncertainty around the mean of the conditioning information, making hard-conditional and unconditional forecasts special cases. An application to a small open-economy DSGE model shows that the benefits of conditioning depend crucially on the ability of the model to capture the correlation between the conditioning information and the variables of interest.

We propose a methodology for producing forecast densities for economic aggregates based on disaggregate evidence. Our ensemble predictive methodology utilizes a linear mixture of experts framework to combine the forecast densities from potentially many component models. Each component represents the univariate dynamic process followed by a single disaggregate variable. The ensemble produced from these components approximates the many unknown relationships between the disaggregates and the aggregate by using time-varying weights on the component forecast densities. In our application, we use the disaggregate ensemble approach to forecast US Personal Consumption Expenditure inflation from 1997Q2 to 2008Q1. Our ensemble combining the evidence from 11 disaggregate series outperforms an aggregate autoregressive benchmark, and an aggregate time-varying parameter specification in density forecasting.

We argue that the next generation of macro modellers at Inflation Targeting central banks should adapt a methodology from the weather forecasting literature known as `ensemble modelling'. In this approach, uncertainty about model specifications (e.g., initial conditions, parameters, and boundary conditions) is explicitly accounted for by constructing ensemble predictive densities from a large number of component models. The components allow the modeller to explore a wide range of uncertainties; and the resulting ensemble `integrates out' these uncertainties using time-varying weights on the components. We provide two examples of this modelling strategy: (i) forecasting inflation with a disaggregate ensemble; and (ii) forecasting inflation with an ensemble DSGE.

An approximate dynamic factor model can substantially improve the reliability of real time output gap estimates. The model extracts a common component from macroeconomic indicators, which reduces errors in the gap due to data revisions. The model’s ability to handle the unbalanced arrival of data, also yields favorable nowcasting properties and thus starting conditions for the filtering of data into trend and deviations from trend. Combined with the method of augmenting data with forecasts prior to filtering, this greatly reduces the end-of-sample imprecision in the gap estimate. The increased precision has economic significance for real time policy decisions.

We investigate optimal horizons for targeting inflation in response to different shocks and their properties under alternative preferences of an inflation-targeting central bank. Our analysis is based on a well specified macroeconometric model of Norway, but we examine how alternative specifications of its key equations would affect our results. We find that the optimal horizon is highly shock-specific, precluding general conclusions for demand and supply shocks. An extension of the horizon with concern for output and/or interest rate fluctuations beyond some shock-specific level proves counterproductive. The size of a given shock does not affect the horizon unless the central bank cares about interest rate volatility, while its sign does not matter unless the model is non-linear. The optimal horizon in response to a combination of shocks cannot be derived from those for each of the shocks, as different shocks may amplify or modify the effects of each other. In this case, however, sources of shocks as well as their sizes and signs become relevant, leading to complex dynamics of inflation and output. Successful inflation targeting in such cases may require a complex interest rate response. The optimal horizon generally increases with the degree of persistence in a shock and decreases with the strength of stabilisation mechanisms in the model.

We investigate empirically whether a central bank can promote financial stability by stabilizing inflation and output, and whether additional stabilization of asset prices and credit growth would enhance financial stability in particular. We employ an econometric model of the Norwegian economy to investigate the performance of simple interest rate rules that allow a response to asset prices and credit growth, in addition to inflation and output. We find that output stabilization tends to improve financial stability. Additional stabilization of house prices, equity prices and/or credit growth enhances stability in both inflation and output, but has mixed effects on financial stability. In general, financial stability as measured by e.g. asset price volatility improves, while financial stability measured by indicators that depend directly on interest rates deteriorates, mainly because of higher interest rate volatility owing to a more active monetary policy.

It is often argued that Norway’s sizeable net foreign assets based on its petroleum wealth imply an appreciation of its real exchange rate to a permanently strong level. We investigate this issue within the framework of the fundamental equilibrium real exchange rate (FEER) approach. It is shown that the strength of the FEER depends on the share of imports that can be financed by petroleum (based) revenues. Projections of the FEER over a long horizon suggest that the petroleum wealth implies a stronger equilibrium exchange rate than the rate that would have balanced (non-petroleum) foreign trade in each period. However, the FEER depreciates steadily over time with growth in imports relative to petroleum revenues and converges towards the rate that balances foreign trade. A permanently strong FEER presupposes that e.g. imports stay constant over time. Our results are in accord with the behaviour of the real exchange before and after the discovery of Norway’s petroleum resources.

By this thesis Eivind Bernhardsen completed his cand.oecon. degree at the University of Oslo. The thesis was financed by Norges Bank and was written as a part of a project on the development of supervisory credit risk models. In this thesis, a model of bankruptcy prediction conditional on financial statements is presented. Apart from giving a discussion on the suggested variables the issue of functional form is raised. The specification most commonly applied for the bankruptcy prediction model implies that the rate at which two variables can substitute another holding predicted risk unchanged will be constant. If the aspect captured by single financial ratios is considered less a substitute for any other aspect as this ratio grows, this restriction may not be appropriate. Specifically, the structure of constant compensation will make predictions sensitive to non-credible outliers. A specification of the logit model which allows for flexible rates of compensation is motivated. The model is estimated and the regression results are reported. Second; by questioning the direct connections between financial ratios and the particular outcome of bankruptcy, a model structure which determines an upper bound on probability estimates is explored. By reference to a simple model of misclassification, the specification distinguishes between the probability of bankruptcy and the probability of insolvency. Whereas the predicted probabilities of bankruptcy can be evaluated empirically, the event of insolvency is not observable. Nevertheless; conditional on the model structure, probabilities can be derived for this event as well. An evaluation is given on the ability of the model to measure the over-all development in credit risk for the Norwegian limited liability sector. Individual probabilities of bankruptcy are multiplied with the firms debt to generate a prediction of expected loss in absence of recovered values. This measure is then aggregated and fitted with total loan losses for the Norwegian banking sector over the years 1989-2000. Finally, the possibility of assessing the effect of macro variables in a short panel of firms is explored. With reference to an aggregation property of the probit model, a suggestion is given on how to estimate time-specific effects on aggregate data as a means to identify macro coefficients that can be included in the micro-level model.

Recent work by Clements and Hendry have shown why forecasting systems that are in terms of differences, dVARs, can be more accurate than economet- ric models that include levels variables, ECMs. For example, dVAR forecasts are insulated from parameter non-constancies in the long run mean of the cointegration relationships. In this paper, the practical relevance of these is- sues are investigated for RIMINI, the quarterly macroeconometric model used in Norges Bank (The Central Bank of Norway), which we take as an example of an ECM forecasting model. We develop two dVAR versions of the full RIMINI model and compare ECM and dVAR forecasts for the period 1992.1- 1994.4. In addition we compare forecasts from the full scale models with those of univariate dVAR type models. The results seems to confirm the relevance of several important theoretical insights. dVAR forecasts appear to provide some immunity against parameter non-constancies that could seriously bias the ECM forecasts. On the other hand however, for open systems like the RI- MINI model, the misspecification resulting from omitting levels information seems to generate substantial biases in the dVAR forecasts. Therefore, the incumbent ECM performs comparatively well over the forecast period inves- tigated in this paper.