|
on Forecasting |
| By: | MacDonald, Stephen; Ash, Mark; Cooke, Bryce |
| Abstract: | We derive a set of stylized facts about USDA’s soybean supply and demand forecasts and draw implications from these results for efforts to improve the accuracy of these forecasts. USDA’s short run soybean supply and demand forecasts are inefficient, with several key variables significantly biased throughout much of the annual forecast cycle. Bias and other characteristics have varied significantly over recent decades, and evaluation efforts intended to guide forecast improvement need to focus on very recent sample years and seasonally disaggregated data. USDA should expand the information set used in its forecasting of U.S. soybean yields, take note of its mid-season downward bias in its Brazilian production forecasts, and shift the weight of its early-season focus for China’s soybean consumption estimates towards forecasts of growth rates and reduce the weight it applies to forecasting consumption and trade by China in terms of levels. During 2004-2015, downward-biased U.S. production forecasts and upward-biased foreign excess supply forecasts resulted in downward-biased U.S. export forecasts throughout much of USDA’s annual forecasting cycle. Twenty years earlier, this downward bias was confined to the end of the forecasting cycle, whereas U.S. soybean ending stock forecasts have been upward biased for decades across much of the forecasting cycle. The forecasts for U.S. soybean exports are also characterized by smoothing, with strong correlation between month-to-month forecast revisions towards the later months of USDA’s annual forecasting cycle. |
| Keywords: | Forecast evaluation, overlapping forecasts, bias, smoothing, agriculture, soybeans, China, United States, Brazil, Argentina, USDA, WASDE |
| JEL: | C53 Q02 Q11 Q17 |
| Date: | 2017–09 |
| URL: | http://d.repec.org/n?u=RePEc:pra:mprapa:87545&r=for |
| By: | Mari Tanaka; Nicholas Bloom; Joel M. David; Maiko Koga |
| Abstract: | Ever since Keynes’ famous quote about animal spirits, there has been an interest in linking firms’ expectations and actions. However, empirical evidence has been limited due to a lack of firm-level panel data on expectations and outcomes. In this paper, we build such a dataset by combining a unique survey of Japanese firms’ GDP forecasts with company accounting data for 25 years for over 1,000 large Japanese firms. We find four main results. First, firms’ GDP forecasts are positively associated with their employment, investment, and output growth in the subsequent year. Second, both optimistic and pessimistic forecast errors lower profitability. Third, while over-optimistic forecasts lower measured productivity, over-pessimistic forecasts do not tend to have an effect on productivity. Overall, these results are stronger for firms whose performance is more sensitive to the state of macroeconomy. We show that a simple model of firm input choice under uncertainty and costly adjustment can rationalize there results. Finally, larger and more cyclically sensitive firms make more accurate forecasts, presumably reflecting a higher return to accuracy for these firms. More productive, older, and bank-owned firms also make more accurate forecasts, suggesting that forecasting ability is also linked to management ability, experience, and governance. Collectively, our results highlight the importance of firms’ forecasting ability for micro and macro performance. |
| JEL: | E0 M0 |
| Date: | 2018–06 |
| URL: | http://d.repec.org/n?u=RePEc:nbr:nberwo:24776&r=for |
| By: | Hyeong Kyu Choi |
| Abstract: | Predicting the price correlation of two assets for future time periods is important in portfolio optimization. We apply LSTM recurrent neural networks (RNN) in predicting the stock price correlation coefficient of two individual stocks. RNNs are competent in understanding temporal dependencies. The use of LSTM cells further enhances its long term predictive properties. To encompass both linearity and nonlinearity in the model, we adopt the ARIMA model as well. The ARIMA model filters linear tendencies in the data and passes on the residual value to the LSTM model. The ARIMA LSTM hybrid model is tested against other traditional predictive financial models such as the full historical model, constant correlation model, single index model and the multi group model. In our empirical study, the predictive ability of the ARIMA-LSTM model turned out superior to all other financial models by a significant scale. Our work implies that it is worth considering the ARIMA LSTM model to forecast correlation coefficient for portfolio optimization. |
| Date: | 2018–08 |
| URL: | http://d.repec.org/n?u=RePEc:arx:papers:1808.01560&r=for |
| By: | Jarek Duda; Ma{\l}gorzata Snarska |
| Abstract: | US Yield curve has recently collapsed to its most flattened level since subprime crisis and is close to the inversion. This fact has gathered attention of investors around the world and revived the discussion of proper modeling and forecasting yield curve, since changes in interest rate structure are believed to represent investors expectations about the future state of economy and have foreshadowed recessions in the United States. While changes in term structure of interest rates are relatively easy to interpret they are however very difficult to model and forecast due to no proper economic theory underlying such events. Yield curves are usually represented by multivariate sparse time series, at any point in time infinite dimensional curve is portrayed via relatively few points in a multivariate space of data and as a consequence multimodal statistical dependencies behind these curves are relatively hard to extract and forecast via typical multivariate statistical methods.We propose to model yield curves via reconstruction of joint probability distribution of parameters in functional space as a high degree polynomial. Thanks to adoption of an orthonormal basis, the MSE estimation of coefficients of a given function is an average over a data sample in the space of functions. Since such polynomial coefficients are independent and have cumulant-like interpretation ie.describe corresponding perturbation from an uniform joint distribution, our approach can also be extended to any d-dimensional space of yield curve parameters (also in neighboring times) due to controllable accuracy. We believe that this approach to modeling of local behavior of a sparse multivariate curved time series can complement prediction from standard models like ARIMA, that are using long range dependencies, but provide only inaccurate prediction of probability distribution, often as just Gaussian with constant width. |
| Date: | 2018–07 |
| URL: | http://d.repec.org/n?u=RePEc:arx:papers:1807.11743&r=for |
| By: | Ray C. Fair (Cowles Foundation, Yale University) |
| Abstract: | This paper examines the question whether information is contained in forecasts from DSGE models beyond that contained in lagged values, which are extensively used in the models. Four sets of forecasts are examined. The results are encouraging for DSGE forecasts of real GDP. The results suggest that there is information in the DSGE forecasts not contained in forecasts based only on lagged values and that there is no information in the lagged-value forecasts not contained in the DSGE forecasts. The opposite is true for forecasts of the GDP deflator. |
| Keywords: | DSGE forecasts, Lagged values |
| JEL: | E10 E17 C53 |
| Date: | 2018–08 |
| URL: | http://d.repec.org/n?u=RePEc:cwl:cwldpp:2140&r=for |