Following on a prior question Objective evidence of USA economic "boom", I took a look at the GDP forecasts for 2018 Q2 provided by the Federal Reserve Bank of Atlanta and NY. As of May 9 the Atlanta Fed is forecasting a 4.0 annual rate for Q2. In contrast the NY Fed is forecasting (as of May 11) a 2.97 Q2 rate. Atlanta here and New York here

(Presumably both banks have the same input data available)

So I wanted to find out how accurate these forecasts have been historically. Unfortunately I can't locate any historical Atlanta forecasts, but I was able to find past forecasts made NY: enter image description here

Unless I am misreading this graph, the NY Fed (on Dec 29 2017) was anticipating a 3.87% real growth GDP. A month later the actual rate was announced as 2.53% and three months later (Mar 29th) revised to 2.89% - both actuals substantially removed from the predicted 3.87.

Again, I may be misinterpreting, (if so I'd hope that a respondent can correct me) but based on limited information is would seem that the NY Fed Reserve forecasts are inaccurate.

Any information that respondents might be able to provide to judge the Atlanta Fed accuracy would be appreciated.

UPDATE to this question to this question: Quoted from Atlanta Fed Macroblog May 22 2017

Indeed, current quarter GDP forecasting models maintained by the Federal Reserve Banks of New York, St. Louis, and Atlanta have been pointing toward stronger second quarter growth (2.3 percent, 2.6 percent and 4.1 percent, as reported on their respective websites on May 19, 2017).

Assuming that the three Federal Reserve banks all have the same current economic data available, how does someone explain the substantial differences between these forecasts.

But I want to maintain focus on this one question: From a historical perspective, how well have these Federal Reserve forecasts predicted the official (BEA determined) GDPs for previous quarters?

  • 1
    $\begingroup$ accurate or inaccurate by what measure? Do these places really only provide point forecasts, or do they actually provide fan forecasts? $\endgroup$ – EnergyNumbers May 14 '18 at 19:17
  • $\begingroup$ Accuracy as defined by how close the forecast is compared with actual. It appears that the forecast example above is "off by 25%". I don't know what you mean by a "fan" forecast. Do you mean a range? Did you check the links? $\endgroup$ – BobE May 14 '18 at 19:30
  • 1
    $\begingroup$ If you look at the description, these are actually “nowcasts” - what is the estimate for GDP based on available data. This methodology is going to have a hard time with revisisions, since it is based on current data. Also, there’s a lot of descriptive detail provided on the Atlanta Fed site, which might cover this. $\endgroup$ – Brian Romanchuk May 14 '18 at 22:33
  • $\begingroup$ the descriptive video on the Atlanta site states: "GDPnow is the Atlanta Fed's forecasting model for the next estimate of GDP growth estimated by the BEA.. it is designed to converge close to what the official (BEA) estimate's going to be shortly before the (actual) BEA release". ... Since it is a forecast how well does that forecast match up with the official BEA released results. So I understand that these are periodic (generally one week periods) forecasts made throughout the quarter. How accurately do these forecasts (aka "nowcasts") "converge" to the official BEA releases? $\endgroup$ – BobE May 15 '18 at 0:41

Your Answer

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

Browse other questions tagged or ask your own question.