Error Statistics for the Survey of Professional Forecasters for Real GNP/GDP Release Date: 09/01/2009 Tom Stark Assistant Director and Manager Real-Time Data Research Center Economic Research Department Federal Reserve Bank of Philadelphia 1. OVERVIEW. This document reports error statistics for median projections from the Survey of Professional Forecasters (SPF), conducted since 1990 by the Federal Reserve Bank of Philadelphia. We provide the results in a series of tables and, in the PDF version of this document, a number of charts. The tables show the survey variable forecast and, importantly, the transformation of the data that we used to generate the statistics. (The transformation is usually a quarter-over-quarter growth rate, expressed in annualized percentage points. However, some variables, such as interest rates, the unemployment rate, and housing starts are untransformed and, thus, expressed in their natural units.) The paragraphs below explain the format of the tables and charts and the methods used to compute the statistics. These paragraphs are general. The same discussion applies to all variables in the survey. 2. DESCRIPTION OF TABLES. Table 1 reports error statistics for various forecast horizons, sample periods, and choices of the real-time historical value that we used to assess accuracy. In each quarterly survey, we ask our panelists for their projections for the current quarter and the next four quarters. The current quarter is defined as the quarter in which we conducted the survey. Our tables provide error statistics separately for each quarter of this five-quarter horizon, beginning with the current quarter (denoted H = 1) and ending with the quarter that is four quarters in the future (H = 5). For each horizon, we report the mean forecast error [ME(S)], the mean absolute forecast error [MAE(S)], and the root-mean-square error [RMSE(S)]. All are standard measures of accuracy, though the academic literature generally places the most weight on the latter. We define a forecast error as the difference between the historical value and the forecast. The mean error for each horizon is simply the average of the forecast errors at that horizon, constructed over the sample period shown in Table 1. Other things the same, a forecast with a mean error close to zero is better than one with a mean error far from zero. The mean absolute error is the sample average of the absolute value of the errors. Many analysts prefer this measure to the mean error because it does not allow large positive errors to offset large negative errors. In this sense, the mean absolute error gives a cleaner estimate of the size of the errors. Decision makers, however, may care not only about the average size of the errors but also about their variability, as measured by variance. Our last measure of accuracy is one that reflects the influence of the mean error and the variance of the error. The root-mean-square error for the SPF [RMSE(S)], the measure most often used by analysts and academicians, is the square root of the the average squared error. The lower the root-mean-square error, the more accurate the forecast. 2.1. Benchmark Models. The forecast error statistics from the SPF are of interest in their own right. However, it is often more interesting to compare such statistics with those of alternative, or benchmark, forecasts. Table 1 reports four such comparisons. It shows the ratio of the root-mean-square error of the SPF forecast to that of four benchmark models. The benchmark models are statistical equations that we estimate on the data. We use the equations to generate projections for the same horizons included in the survey. In effect, we imagine standing back in time at each date when a survey was conducted and generating a separate forecast with each benchmark model. We do this in the same way that a survey panelist would have done using his own model. A RMSE ratio below unity indicates that the SPF consensus (median) forecast has a root-mean-square error lower than that of the benchmark. This means the SPF is more accurate. We now describe the benchmark models. The first is perhaps the simplest of all possible benchmarks: A no-change model. In this model, the forecast for quarter T, the one-step-ahead or current-quarter forecast, is simply the historical value for the prior quarter (T - 1). There is, in other words, no change in the forecast compared with the historical value. Moreover, the forecast for the remaining quarters of the horizon is the same as the forecast for the current quarter. We denote the relative RMSE ratio for this benchmark as RMSE(S/NC), using NC to indicate no change. The second and third benchmark models generate projections using one or more historical observations of the the variable forecast, weighted by coefficients estimated from the data. Such autoregressive (AR) models can be formulated in two ways. We can estimate one model to generate the forecasts at all horizons, using an iteration method to generate the projections beyond the current quarter (IAR), or we can directly estimate a new model for each forecast horizon (DAR). The latter formulation has been shown to reduce the bias in a forecast when the underlying model is characterized by certain types of misspecification. The root-mean-square error ratios are denoted RMSE(S/IAR) and RMSE(S/DAR), respectively. The one- through five-step-ahead projections of the benchmark models use information on the quarterly average of the variable forecast. The latest historical observation is for the quarter that is one quarter before the quarter of the first projection in the horizon. In contrast, the panelists generate their projections with the help of additional information. They submit their projections near the middle of each quarter and hence have access to some monthly indicators for the first month of each quarter, when those data are released before the survey deadline. This puts the projections of panelists for some variables at an advantage relative to the corresponding benchmark projections. Moreover, the panelists may also examine the very recent historical values of such monthly indicators in forming their projections for quarterly averages. Such monthly statistical momentum represents an advantage not shared by the benchmark models, which use only quarterly averages. For survey variables whose observations are reported at a monthly frequency, such as interest rates, industrial production, housing starts, and unemployment, we estimate and forecast a fourth benchmark model, the DARM. This model adds recent monthly historical values to the specification of the DAR model. For the projections for unemployment, nonfarm payroll employment, and interest rates, we add the values of monthly observations, beginning with that for the first month of the first quarter of the forecast horizon. These values should be in the information set of the survey panelists at the time they formed their projections. In contrast, for variables such as housing starts and industrial production, we include only lagged values of monthly observations. For such variables, the panelists would not have known the monthly observation for the first month of the first quarter of the forecast horizon. In general, we find that adding monthly observations to the benchmark DAR models improves accuracy. Indeed, for the projections for interest rates and the unemployment rate, the accuracy of the benchmark DARM projections rivals that of the SPF projections. 2.2. Real-Time Data. All benchmark models are estimated on a rolling, fixed window of 60 real-time quarterly observations. Lag lengths, based on either the Akaike information criterion (AIC) or the Schwarz information criterion (SIC), are re-estimated each period. The tables below indicate whether the lag length was was chosen by the AIC or SIC. We would like to make the comparison between the SPF forecast and the forecasts of each benchmark as fair as possible. Therefore, we must subject the benchmark models to the same data environment the survey panelists faced when they made their projections. This is important because macroeconomic data are revised often, and we do not want the benchmark models to use a data set that differs from the one our panelists would have used. We estimate and forecast the benchmark models with real-time data from the Philadelphia Fed real-time data set, using the vintage of data that the survey panelists would have had at the time they generated their own projections. (For more information on the Philadelphia Fed real-time data set, go to www.philadelphiafed.org/econ/forecast/real-time-data/.) An open question in the literature on forecasting is: What version or vintage of the data should we use to compute the errors? A closely related question is: What version of the data are professional forecasters trying to predict? Our computations take no strong position on these questions. Rather, in Table 1, we evaluate the projections (SPF and benchmark) with five alternative measures of the historical values, all from the Philadelphia Fed real-time data set. These measures range from the initial-release values to the values as we know them today. All together, we compute the forecast error statistics using the following five alternative measures of historical values: (1) The initial or first-release value; (2) The revised value as it appears one quarter after the initial release; (3) The revised value as it appears five quarters after the initial release; (4) The revised value as it appears nine quarters after the initial release; (5) The revised value as it appears today. Each measure of the historical value has advantages and disadvantages. The initial-release value is the first measure released by government statistical agencies. A forecaster might be very interested in this measure because it enables him to evaluate his latest forecast soon after he generated it. However, early releases of the data are often subject to large measurement error. Subsequent releases [(2) - (5)] are more accurate, but they are available much later than the initial release. As we go from the first measure to the fifth, we get more reliability, at the cost of higher delays in availability. The last two columns in Table 1 report the number of observations that we used to compute the error statistics. Some observations are omitted because the data are missing in the real-time data set, such as occurred when federal government statistical agencies closed in late 1995. 2.3. Recent Projections and Realizations. Tables 2 to 7 provide information on recent projections and realizations. They show how we align the data prior to computing the forecast errors that form the backbone of the computations in Table 1. Any forecast error can be written as the equation given by error = realization - forecast. For our computations, we must be more precise because, for each projection (SPF and benchmarks), we have different periods forecast (T) different forecast horizons (h), and several measures of the realization (m). Thus, we can define the forecast error more precisely as error( T, h, m ) = realization( T, m ) - forecast( T, h ). Tables 2 to 7 are organized along these lines. Table 2 shows recent forecasts from the SPF. Each column gives the projection for a different horizon or forecast step (h), beginning with that for the current quarter, defined as the quarter in which we conducted the survey. The dates (T) given in the rows show the periods forecast. These also correspond to the dates that we conducted the survey. Tables 3 to 6 report the recent projections of the four benchmark models. These are organized in the same way as Table 2. Table 7 reports recent values of the five alternative realizations (m) we use to compute the error statistics. 2.4. Qualifications. We note two minor qualifications to the methods discussed above. The first concerns the vintage of data that we used to estimate and forecast the benchmark models for CPI inflation. The second concerns the five measures of realizations used for the unemployment rate, nonfarm payroll employment, and CPI inflation. To estimate and forecast the benchmark models for CPI inflation, we use the vintage of data that would have been available in the middle of each quarter. This predates by one month the vintage that SPF panelists would have had at their disposal when they formed their projections. The effect is likely small because revisions to the CPI are generally small. To compute the realizations for unemployment, nonfarm payroll employment, and CPI inflation, we use the vintages associated with the middle of each quarter. The measure that we call initial comes from this vintage, even though the initial estimate was available one month earlier. Thus, for these variables, our initial estimate reflects some revision by government statistical agencies. The effect for unemployment and CPI inflation is likely small. The effect could be somewhat larger for nonfarm payroll employment. 3. DESCRIPTION OF GRAPHS. 3.1. Root-Mean-Square Errors. For each sample period shown in Table 1, we provide graphs of the root-mean-square error for the SPF forecast. There is one page for each sample period. On each page, we plot (for each forecast horizon) the RMSE on the y-axis. The x-axis shows the measure of the historical value that we used to compute the RMSE. These range from the value on its initial release to the value one quarter later to the value as we know it now (at the time we made the computation). The graphs provide a tremendous amount of information. If we focus on a particular graph, we can see how a change in the measure of the realization (x-axis) affects the root-mean-square-error measure of accuracy. The effect is pronounced for some variables, such as real GDP and some of its components. For others, there is little or no effect. For example, because the historical data on interest rates are not revised, the estimated RMSE is the same in each case. If we compare a particular point on one graph with the same point on another, we see how the forecast horizon affects accuracy. In general, the RMSE rises (accuracy falls) as the forecast horizon lengthens. Finally, if we compare a graph on one page with the corresponding graph on another page, we see how our estimates of accuracy in the SPF change with the sample period. Periods characterized by a high degree of economic turbulence will generally produce large RMSEs. 3.2. Fan Charts. The last chart plots recent historical values and the latest SPF forecast. It also shows confidence intervals for the forecast, based on back-of-the-envelope calculations. The historical values and the SPF forecast are those associated with the latest vintage of data and survey, respectively, available at the time we ran our computer programs. The confidence intervals are constructed under the assumption that the historical forecast errors over the sample (shown in the footnote) follow a normal distribution with a mean of zero and a variance given by the root-mean-square error. The latter is estimated over the aforementioned sample, using the measure of history listed in the footnote. ---------------------------------------------------------------------------------- Table 1. Forecast Error Statistics for SPF Variable: RGDP (Real GNP/GDP) ---------------------------------------------------------------------------------- Computed Over Various Sample Periods Various Measures of Realizations Transformation: Q/Q Growth Rate Lag Length for IAR(p), DAR(p), and DARM(p) Models: AIC Last Updated: 09/01/2009 10:22 ---------------------------------------------------------------------------------- H ME(S) MAE(S) RMSE(S) RMSE(S/NC) RMSE(S/IAR) RMSE(S/DAR) RMSE(S/DARM) Nspf N History: Initial Release 1985:01-2007:01 1 0.29 1.14 1.41 0.69 0.81 0.81 NA 88 87 2 0.06 1.32 1.66 0.78 0.93 0.94 NA 88 87 3 0.02 1.45 1.77 0.77 0.98 0.98 NA 88 87 4 -0.09 1.40 1.81 0.77 1.00 1.00 NA 88 87 5 -0.09 1.40 1.79 0.74 0.99 0.96 NA 88 87 1985:01-1996:04 1 0.19 1.00 1.20 0.62 0.72 0.72 NA 47 46 2 -0.11 1.16 1.42 0.73 0.82 0.84 NA 47 46 3 -0.16 1.30 1.59 0.71 0.88 0.88 NA 47 46 4 -0.35 1.21 1.65 0.70 0.91 0.91 NA 47 46 5 -0.27 1.23 1.64 0.71 0.91 0.86 NA 47 47 1997:01-2007:01 1 0.41 1.30 1.62 0.75 0.90 0.90 NA 41 41 2 0.25 1.50 1.90 0.81 1.03 1.03 NA 41 41 3 0.23 1.63 1.96 0.83 1.08 1.09 NA 41 41 4 0.20 1.62 1.98 0.85 1.08 1.09 NA 41 41 5 0.12 1.60 1.94 0.76 1.07 1.07 NA 41 40 H ME(S) MAE(S) RMSE(S) RMSE(S/NC) RMSE(S/IAR) RMSE(S/DAR) RMSE(S/DARM) Nspf N History: One Qtr After Initial Release 1985:01-2007:01 1 0.37 1.31 1.68 0.78 0.88 0.88 NA 89 88 2 0.13 1.47 1.82 0.84 0.93 0.94 NA 89 88 3 0.10 1.53 1.90 0.77 0.95 0.95 NA 89 88 4 -0.01 1.50 1.94 0.76 0.96 0.96 NA 89 88 5 -0.01 1.51 1.96 0.74 0.97 0.95 NA 89 88 1985:01-1996:04 1 0.18 1.20 1.52 0.76 0.83 0.83 NA 48 47 2 -0.12 1.33 1.61 0.85 0.86 0.88 NA 48 47 3 -0.17 1.39 1.68 0.71 0.85 0.85 NA 48 47 4 -0.35 1.31 1.73 0.67 0.88 0.87 NA 48 47 5 -0.28 1.38 1.79 0.70 0.92 0.89 NA 48 48 1997:01-2007:01 1 0.59 1.43 1.85 0.80 0.92 0.92 NA 41 41 2 0.43 1.63 2.05 0.83 1.00 0.99 NA 41 41 3 0.41 1.69 2.14 0.84 1.04 1.05 NA 41 41 4 0.39 1.72 2.16 0.86 1.05 1.04 NA 41 41 5 0.30 1.68 2.13 0.77 1.03 1.02 NA 41 40 H ME(S) MAE(S) RMSE(S) RMSE(S/NC) RMSE(S/IAR) RMSE(S/DAR) RMSE(S/DARM) Nspf N History: Five Qtrs After Initial Release 1985:01-2007:01 1 0.36 1.44 1.82 0.77 0.86 0.86 NA 89 88 2 0.13 1.57 1.98 0.82 0.93 0.94 NA 89 88 3 0.09 1.66 2.08 0.83 0.96 0.97 NA 89 88 4 -0.02 1.61 2.08 0.75 0.95 0.95 NA 89 88 5 -0.02 1.59 2.08 0.75 0.96 0.92 NA 89 88 1985:01-1996:04 1 0.30 1.41 1.73 0.74 0.81 0.81 NA 48 47 2 -0.00 1.53 1.87 0.82 0.88 0.90 NA 48 47 3 -0.05 1.60 1.96 0.81 0.89 0.90 NA 48 47 4 -0.23 1.49 1.98 0.70 0.89 0.89 NA 48 47 5 -0.16 1.50 1.97 0.73 0.90 0.84 NA 48 48 1997:01-2007:01 1 0.44 1.48 1.93 0.79 0.92 0.92 NA 41 41 2 0.28 1.61 2.11 0.83 0.99 0.99 NA 41 41 3 0.25 1.74 2.20 0.85 1.03 1.05 NA 41 41 4 0.23 1.74 2.19 0.81 1.02 1.02 NA 41 41 5 0.15 1.69 2.20 0.76 1.02 1.01 NA 41 40 H ME(S) MAE(S) RMSE(S) RMSE(S/NC) RMSE(S/IAR) RMSE(S/DAR) RMSE(S/DARM) Nspf N History: Nine Qtrs After Initial Release 1985:01-2007:01 1 0.30 1.48 1.89 0.81 0.88 0.88 NA 89 88 2 0.07 1.62 2.06 0.83 0.94 0.94 NA 89 88 3 0.03 1.71 2.18 0.84 0.97 0.97 NA 89 88 4 -0.07 1.66 2.18 0.79 0.97 0.96 NA 89 88 5 -0.08 1.67 2.22 0.78 0.99 0.94 NA 89 88 1985:01-1996:04 1 0.38 1.50 1.90 0.80 0.86 0.86 NA 48 47 2 0.08 1.55 2.01 0.85 0.91 0.92 NA 48 47 3 0.04 1.65 2.10 0.83 0.91 0.91 NA 48 47 4 -0.14 1.54 2.11 0.75 0.91 0.90 NA 48 47 5 -0.07 1.55 2.15 0.78 0.94 0.87 NA 48 48 1997:01-2007:01 1 0.21 1.47 1.89 0.82 0.90 0.90 NA 41 41 2 0.05 1.70 2.12 0.81 0.98 0.97 NA 41 41 3 0.03 1.77 2.26 0.84 1.04 1.05 NA 41 41 4 0.00 1.79 2.26 0.85 1.04 1.03 NA 41 41 5 -0.08 1.81 2.29 0.78 1.04 1.03 NA 41 40 H ME(S) MAE(S) RMSE(S) RMSE(S/NC) RMSE(S/IAR) RMSE(S/DAR) RMSE(S/DARM) Nspf N History: Latest Vintage 1985:01-2007:01 1 0.63 1.57 1.94 0.79 0.91 0.91 NA 89 88 2 0.40 1.65 2.08 0.85 0.99 0.99 NA 89 88 3 0.36 1.70 2.18 0.86 1.00 1.01 NA 89 88 4 0.26 1.68 2.18 0.80 1.00 0.99 NA 89 88 5 0.25 1.70 2.22 0.82 1.01 0.96 NA 89 88 1985:01-1996:04 1 0.87 1.49 1.87 0.76 0.88 0.88 NA 48 47 2 0.57 1.47 1.89 0.82 0.93 0.94 NA 48 47 3 0.53 1.52 1.99 0.80 0.92 0.92 NA 48 47 4 0.35 1.48 1.97 0.72 0.92 0.91 NA 48 47 5 0.42 1.51 2.00 0.76 0.94 0.86 NA 48 48 1997:01-2007:01 1 0.35 1.66 2.03 0.82 0.94 0.94 NA 41 41 2 0.19 1.86 2.29 0.88 1.04 1.04 NA 41 41 3 0.17 1.90 2.39 0.91 1.09 1.09 NA 41 41 4 0.15 1.90 2.40 0.88 1.08 1.07 NA 41 41 5 0.06 1.94 2.45 0.88 1.09 1.07 NA 41 40 Table 1 notes. (1) The forecast horizon is given by H, where H = 1 is the SPF forecast for the current quarter. (2) The headers ME(S), MAE(S), and RMSE(S) are mean error, mean absolute error, and root-mean-square error for the SPF. (3) The header RMSE(S/NC) is the ratio of the SPF RMSE to that of the no-change (NC) model. (4) The headers RMSE(S/IAR), RMSE(S/DAR) and RMSE(S/DARM) are the ratios of the SPF RMSE to the RMSE of the iterated and direct autoregressive models and the direct autoregressive model augmented with monthly observations, respectively. All models are estimated on a rolling window of 60 observations from the Phila Fed real-time data set. (5) The headers Nspf and N are the number of observations analyzed for the SPF and benchmark models. (6) When the variable forecast is a growth rate or an interest rate, it is expressed in annualized percentage points. When the variable forecast is the unemployment rate, it is expressed in percentage points. (7) Sample periods refer to the dates forecast, not the dates when the forecasts were made. Source: Tom Stark, Research Department, FRB Philadelphia. -------------------------------------------------------------------- Table 2. Recent SPF Forecasts - Dated at the Quarter Forecast -------------------------------------------------------------------- Variable: RGDP (Real GNP/GDP) By Forecast Step (1 to 5) Transformation: Q/Q Growth Rate Last Updated: 09/01/2009 10:22 -------------------------------------------------------------------- Qtr Forecast Step 1 Step 2 Step 3 Step 4 Step 5 2002:04 1.330 2.634 3.416 3.239 4.007 2003:01 2.169 2.566 3.439 3.501 3.354 2003:02 1.800 2.735 3.052 3.051 3.612 2003:03 3.537 3.364 3.434 3.277 3.657 2003:04 4.000 3.863 3.356 3.561 4.165 2004:01 4.449 3.949 3.850 4.015 3.824 2004:02 4.497 4.276 3.875 3.848 3.407 2004:03 3.485 4.128 3.939 3.675 3.598 2004:04 3.700 4.015 4.013 4.082 3.785 2005:01 3.717 3.367 3.764 3.801 3.649 2005:02 3.000 3.665 3.433 3.694 3.955 2005:03 4.200 3.501 3.315 3.430 3.343 2005:04 3.200 3.600 3.434 3.390 3.367 2006:01 4.397 3.724 3.258 3.233 3.573 2006:02 3.400 3.353 3.271 3.355 3.379 2006:03 2.727 3.100 3.001 3.204 3.253 2006:04 2.484 2.872 3.000 3.242 3.199 2007:01 2.566 2.716 2.870 2.888 3.283 2007:02 2.400 2.734 2.856 2.729 3.024 2007:03 2.500 2.600 3.000 2.945 2.970 2007:04 1.532 2.663 2.864 3.177 2.900 2008:01 0.673 2.153 2.679 2.913 3.062 2008:02 0.205 1.326 2.261 2.920 3.001 2008:03 1.154 1.700 2.759 2.762 2.668 2008:04 -2.935 0.711 1.837 2.819 2.795 2009:01 -5.189 -1.097 1.575 2.271 3.123 2009:02 -1.528 -1.750 0.837 2.111 2.522 2009:03 2.439 0.449 0.973 0.863 2.537 2009:04 NA 2.174 1.726 1.773 2.273 2010:01 NA NA 2.525 2.225 2.439 2010:02 NA NA NA 2.812 2.923 2010:03 NA NA NA NA 2.578 Table 2 notes. (1) Each column gives the sequence of SPF projections for a given forecast step. The forecast steps range from one (the forecast for the quarter in which the survey was conducted) to four quarters in the future (step 5). (2) The dates listed in the rows are the dates forecast, not the dates when the forecasts were made, with the exception of the forecast at step one, for which the two dates coincide. Source: Tom Stark, Research Department, FRB Philadelphia. ------------------------------------------------------------------------------------ Table 3. Recent Benchmark Model 1 IAR Forecasts - Dated at the Quarter Forecast ------------------------------------------------------------------------------------ Variable: RGDP (Real GNP/GDP) By Forecast Step (1 to 5) Transformation: Q/Q Growth Rate Lag Length for IAR(p): AIC Last Updated: 09/01/2009 10:22 ------------------------------------------------------------------------------------ Qtr Forecast Step 1 Step 2 Step 3 Step 4 Step 5 2002:04 2.574 2.768 3.165 2.454 2.758 2003:01 2.669 2.904 2.890 3.134 2.626 2003:02 2.183 2.253 2.842 2.927 3.127 2003:03 2.385 2.374 2.647 2.908 2.938 2003:04 4.146 2.594 2.557 2.625 2.908 2004:01 4.591 4.065 2.649 2.646 2.719 2004:02 3.535 3.632 3.487 2.708 2.710 2004:03 3.445 3.400 3.562 3.318 2.734 2004:04 3.253 3.104 3.211 3.296 3.166 2005:01 3.373 3.277 3.149 3.134 3.217 2005:02 3.280 3.166 3.145 3.058 3.068 2005:03 3.356 3.101 3.189 3.123 3.051 2005:04 3.349 3.229 3.116 3.132 3.079 2006:01 3.189 3.412 3.182 3.067 3.127 2006:02 2.951 2.591 3.279 3.133 3.059 2006:03 3.942 3.772 3.127 3.271 3.108 2006:04 2.890 3.079 3.271 3.028 3.227 2007:01 2.904 2.715 3.478 3.482 3.175 2007:02 2.767 3.304 3.081 3.239 3.347 2007:03 2.318 2.535 3.167 3.061 3.345 2007:04 3.382 3.101 2.982 3.274 3.167 2008:01 3.386 3.360 2.833 2.951 3.242 2008:02 2.133 2.408 3.208 3.056 3.087 2008:03 2.399 2.312 3.088 3.188 2.990 2008:04 2.611 2.694 2.724 2.868 3.142 2009:01 0.563 2.161 2.846 2.816 3.038 2009:02 -3.560 0.635 2.812 2.932 2.931 2009:03 -2.141 -2.505 1.742 2.770 2.977 2009:04 NA -0.401 -1.340 1.987 2.925 2010:01 NA NA -0.300 -0.581 2.314 2010:02 NA NA NA 0.402 0.054 2010:03 NA NA NA NA 0.657 Table 3 notes. (1) Each column gives the sequence of benchmark IAR projections for a given forecast step. The forecast steps range from one to five. The first step corresponds to the forecast that SPF panelists make for the quarter in which the survey is conducted. (2) The dates listed in the rows are the dates forecast, not the dates when the forecasts were made, with the exception of the forecast at step one, for which the two dates coincide. (3) The IAR benchmark model is estimated on a fixed 60-quarter rolling window. Its forecasts are computed with the indirect method. Estimation uses data from the Philadelphia Fed real-time data set. Source: Tom Stark, Research Department, FRB Philadelphia. --------------------------------------------------------------------------------------- Table 4. Recent Benchmark Model 2 No-Change Forecasts - Dated at the Quarter Forecast --------------------------------------------------------------------------------------- Variable: RGDP (Real GNP/GDP) By Forecast Step (1 to 5) Transformation: Q/Q Growth Rate Last Updated: 09/01/2009 10:22 --------------------------------------------------------------------------------------- Qtr Forecast Step 1 Step 2 Step 3 Step 4 Step 5 2002:04 3.137 1.059 5.836 0.224 -0.355 2003:01 0.744 3.137 1.059 5.836 0.224 2003:02 1.598 0.744 3.137 1.059 5.836 2003:03 2.370 1.598 0.744 3.137 1.059 2003:04 7.155 2.370 1.598 0.744 3.137 2004:01 4.024 7.155 2.370 1.598 0.744 2004:02 4.158 4.024 7.155 2.370 1.598 2004:03 3.044 4.158 4.024 7.155 2.370 2004:04 3.711 3.044 4.158 4.024 7.155 2005:01 3.147 3.711 3.044 4.158 4.024 2005:02 3.088 3.147 3.711 3.044 4.158 2005:03 3.414 3.088 3.147 3.711 3.044 2005:04 3.805 3.414 3.088 3.147 3.711 2006:01 1.119 3.805 3.414 3.088 3.147 2006:02 4.818 1.119 3.805 3.414 3.088 2006:03 2.458 4.818 1.119 3.805 3.414 2006:04 1.583 2.458 4.818 1.119 3.805 2007:01 3.473 1.583 2.458 4.818 1.119 2007:02 1.260 3.473 1.583 2.458 4.818 2007:03 3.382 1.260 3.473 1.583 2.458 2007:04 3.896 3.382 1.260 3.473 1.583 2008:01 0.636 3.896 3.382 1.260 3.473 2008:02 0.597 0.636 3.896 3.382 1.260 2008:03 1.889 0.597 0.636 3.896 3.382 2008:04 -0.252 1.889 0.597 0.636 3.896 2009:01 -3.804 -0.252 1.889 0.597 0.636 2009:02 -6.144 -3.804 -0.252 1.889 0.597 2009:03 -1.017 -6.144 -3.804 -0.252 1.889 2009:04 NA -1.017 -6.144 -3.804 -0.252 2010:01 NA NA -1.017 -6.144 -3.804 2010:02 NA NA NA -1.017 -6.144 2010:03 NA NA NA NA -1.017 Table 4 notes. (1) Each column gives the sequence of benchmark no-change projections for a given forecast step. The forecast steps range from one to five. The first step corresponds to the forecast that SPF panelists make for the quarter in which the survey is conducted. (2) The dates listed in the rows are the dates forecast, not the dates when the forecasts were made, with the exception of the forecast at step one, for which the two dates coincide. (3) The projections use data from the Philadelphia Fed real-time data set. Source: Tom Stark, Research Department, FRB Philadelphia. ------------------------------------------------------------------------------------ Table 5. Recent Benchmark Model 3 DAR Forecasts - Dated at the Quarter Forecast ------------------------------------------------------------------------------------ Variable: RGDP (Real GNP/GDP) By Forecast Step (1 to 5) Transformation: Q/Q Growth Rate Lag Length for DAR(p): AIC Last Updated: 09/01/2009 10:22 ------------------------------------------------------------------------------------ Qtr Forecast Step 1 Step 2 Step 3 Step 4 Step 5 2002:04 2.574 2.357 3.285 2.604 3.085 2003:01 2.669 3.002 2.793 3.422 2.974 2003:02 2.183 2.179 2.980 2.743 3.102 2003:03 2.385 2.447 2.705 2.983 2.950 2003:04 4.146 2.673 2.749 2.634 2.967 2004:01 4.591 4.069 2.774 2.729 2.865 2004:02 3.535 3.347 3.263 2.770 2.870 2004:03 3.445 3.354 3.025 3.342 2.827 2004:04 3.253 3.007 3.025 3.030 2.846 2005:01 3.373 3.268 2.980 3.038 2.985 2005:02 3.280 3.097 3.029 2.991 2.961 2005:03 3.356 3.053 3.046 3.065 2.977 2005:04 3.349 3.169 3.028 3.058 2.991 2006:01 3.189 3.399 3.043 3.035 3.035 2006:02 2.951 2.481 3.112 3.075 3.020 2006:03 3.942 3.812 3.002 3.186 3.012 2006:04 2.890 3.022 3.354 2.979 3.071 2007:01 2.904 2.727 3.203 3.507 3.168 2007:02 2.767 3.330 3.179 3.169 3.191 2007:03 2.318 2.558 3.257 3.077 3.252 2007:04 3.382 3.170 3.194 3.288 3.308 2008:01 3.386 3.345 3.101 2.995 3.223 2008:02 2.133 2.371 3.103 3.125 3.294 2008:03 2.399 2.389 3.031 3.191 3.098 2008:04 2.611 2.747 3.026 2.741 3.097 2009:01 0.563 2.158 2.680 2.650 3.023 2009:02 -3.560 0.939 2.824 2.829 3.029 2009:03 -2.141 -1.016 1.321 2.493 2.957 2009:04 NA 0.512 -1.765 1.334 2.913 2010:01 NA NA -1.661 -0.601 3.067 2010:02 NA NA NA 0.990 2.056 2010:03 NA NA NA NA 2.105 Table 5 notes. (1) Each column gives the sequence of benchmark DAR projections for a given forecast step. The forecast steps range from one to five. The first step corresponds to the forecast that SPF panelists make for the quarter in which the survey is conducted. (2) The dates listed in the rows are the dates forecast, not the dates when the forecasts were made, with the exception of the forecast at step one, for which the two dates coincide. (3) The DAR benchmark model is estimated on a fixed 60-quarter rolling window. Its forecasts are computed with the direct method. Estimation uses data from the Philadelphia Fed real-time data set. Source: Tom Stark, Research Department, FRB Philadelphia. ------------------------------------------------------------------------------------ Table 6. Recent Benchmark Model 4 DARM Forecasts - Dated at the Quarter Forecast ------------------------------------------------------------------------------------ Variable: RGDP (Real GNP/GDP) By Forecast Step (1 to 5) Transformation: Q/Q Growth Rate Lag Length for DARM(p): AIC Last Updated: 09/01/2009 10:22 ------------------------------------------------------------------------------------ Qtr Forecast Step 1 Step 2 Step 3 Step 4 Step 5 2002:04 NA NA NA NA NA 2003:01 NA NA NA NA NA 2003:02 NA NA NA NA NA 2003:03 NA NA NA NA NA 2003:04 NA NA NA NA NA 2004:01 NA NA NA NA NA 2004:02 NA NA NA NA NA 2004:03 NA NA NA NA NA 2004:04 NA NA NA NA NA 2005:01 NA NA NA NA NA 2005:02 NA NA NA NA NA 2005:03 NA NA NA NA NA 2005:04 NA NA NA NA NA 2006:01 NA NA NA NA NA 2006:02 NA NA NA NA NA 2006:03 NA NA NA NA NA 2006:04 NA NA NA NA NA 2007:01 NA NA NA NA NA 2007:02 NA NA NA NA NA 2007:03 NA NA NA NA NA 2007:04 NA NA NA NA NA 2008:01 NA NA NA NA NA 2008:02 NA NA NA NA NA 2008:03 NA NA NA NA NA 2008:04 NA NA NA NA NA 2009:01 NA NA NA NA NA 2009:02 NA NA NA NA NA 2009:03 NA NA NA NA NA 2009:04 NA NA NA NA NA 2010:01 NA NA NA NA NA 2010:02 NA NA NA NA NA 2010:03 NA NA NA NA NA Table 6 notes. (1) Each column gives the sequence of benchmark DARM projections for a given forecast step. The forecast steps range from one to five. The first step corresponds to the forecast that SPF panelists make for the quarter in which the survey is conducted. (2) The dates listed in the rows are the dates forecast, not the dates when the forecasts were made, with the exception of the forecast at step one, for which the two dates coincide. (3) The DARM benchmark model is estimated on a fixed 60-quarter rolling window. Its forecasts are computed with the direct method and incorporate recent monthly values of the dependent variable. Estimation uses data from the Philadelphia Fed real-time data set. Source: Tom Stark, Research Department, FRB Philadelphia. ------------------------------------------------------------ Table 7. Recent Realizations (Various Measures) Source: Philadelphia Fed Real-Time Data Set ------------------------------------------------------------ Variable: RGDP (Real GNP/GDP) Transformation: Q/Q Growth Rate Last Updated: 09/01/2009 10:22 1- Initial Release 2- One Qtr After Initial Release 3- Five Qtrs After Initial Release 4- Nine Qtrs After Initial Release 5- Latest Vintage ------------------------------------------------------------- Obs. Date (1) (2) (3) (4) (5) 2002:04 0.744 1.382 1.286 0.737 0.083 2003:01 1.598 1.428 1.928 1.707 1.631 2003:02 2.370 3.281 4.107 3.671 3.229 2003:03 7.155 8.203 7.406 7.246 6.877 2003:04 4.024 4.142 4.185 3.570 3.645 2004:01 4.158 4.489 4.252 3.849 2.847 2004:02 3.044 3.301 3.497 4.042 2.874 2004:03 3.711 4.001 3.974 3.102 2.970 2004:04 3.147 3.848 3.304 2.614 3.518 2005:01 3.088 3.805 3.402 3.073 4.050 2005:02 3.414 3.310 3.264 2.813 1.715 2005:03 3.805 4.142 4.183 4.462 3.076 2005:04 1.119 1.653 1.764 1.194 2.082 2006:01 4.818 5.581 4.820 4.819 5.350 2006:02 2.458 2.559 2.442 2.680 1.451 2006:03 1.583 1.960 1.066 0.796 0.105 2006:04 3.473 2.452 2.091 1.504 2.957 2007:01 1.260 0.602 0.049 1.206 1.206 2007:02 3.382 3.821 4.789 NA 3.217 2007:03 3.896 4.907 4.757 NA 3.595 2007:04 0.636 0.578 -0.172 NA 2.122 2008:01 0.597 0.874 -0.724 NA -0.724 2008:02 1.889 2.825 NA NA 1.456 2008:03 -0.252 -0.511 NA NA -2.677 2008:04 -3.804 -6.342 NA NA -5.373 2009:01 -6.144 -6.429 NA NA -6.429 2009:02 -1.017 NA NA NA -1.017 2009:03 NA NA NA NA NA 2009:04 NA NA NA NA NA 2010:01 NA NA NA NA NA 2010:02 NA NA NA NA NA 2010:03 NA NA NA NA NA Table 7 notes. (1) Each column reports a sequence of realizations from the Philadelphia Fed real-time data set. (2) The date listed in each row is the observation date. (3) Moving across a particular row shows how the observation is revised in subsequent releases. Source: Tom Stark, Research Department, FRB Philadelphia.