The Second of Two Sea Ice Trilogies: Real Time

The second trilogy, below, this time without Rudebusch, also treats Arctic sea ice forecasting, but from a real-time, fixed-target (September), perspective. All of the work came under the umbrella of the Arctic Research Consortium of the United States (ARCUS), and its Sea Ice Prediction Network (SIPN), which oversaw the Sea Ice Outlook (SIO). 

More on ARCUS in the next post. But for now let's look at the second trilogy.

The basic feature-engineered real-time forecasting approach is proposed in

Diebold, F.X. and Gobel, M. (2022), “A Benchmark Model for Fixed-Target Arctic Sea Ice Forecasting,” Economics Letters, 215, 110478,

and it is compared to a feature-engineered real-time machine learning approach in 

Diebold, F.X., Goebel, M., and Goulet Coulombe, P. (2023), “Assessing and Comparing Fixed-Target Forecasts of Arctic Sea Ice: Glide Charts for Feature-Engineered Linear Regression and Machine Learning Models,” Energy Economics, 124, 106833.

The trilogy culminates with a 61-author (must be real science!) assessment of June-September real-time Arctic sea ice forecasting across many models and years, including the above Diebold et al approach: 

Bushuk, M., et al. (2024), “Predicting September Arctic Sea Ice: A Multi-Model Seasonal Skill Comparison,” Bulletin of the American Meteorological Society, 105, 1170-1203. 

The First of Two Arctic Sea Ice Trilogies

I never bogged on the Arctic sea ice "trilogy" below with Glenn Rudebusch et al., certainly not for lack of interest, but rather because much of the research was done when the blog was dormant in the early 2020s. The trilogy addresses the timing of the first ice-free Arctic (IFA) September -- the early IFA (late 2030s) robustly predicted by statistical time-series models; why, in contrast, the large-scale structural climate models tend to get things wrong with a much later IFA; and why you should care. I'm posting on it now not only because I simply think it's interesting and important (and I hope you will too), but also because it complements and contrasts with my next post (on a related but different Arctic sea ice trilogy). Stay tuned! 

Diebold, F.X. and Rudebusch, G.D. (2022), “Probability Assessments of an Ice-Free Arctic: Comparing Statistical and Climate Model Projections,” Journal of Econometrics, 231, 520-534.
---> Compares statistical and large-scale climate model forecasts, and finds that the stat models forecast a much earlier near-ice-free Arctic.

Diebold, F.X., Rudebusch, G.D., Goebel, M., Goulet Coulombe, P. and Zhang, B. (2023), “When Will Arctic Sea Ice Disappear? Projections of Area, Extent, Thickness, and Volume," Journal of Econometrics, 236, 105479.
---> Drills down much deeper on the DR (2022) statistical models, exploring many variations (extent, area, thickness, volume; polynomial vs carbon trends; much more...), establishing robustness of the DR (2022) results and providing more refined forecasts.

Diebold, F.X. and Rudebusch, G.D. (2023), “Climate Models Underestimate the Sensitivity of Arctic Sea Ice to Carbon Emissions,” Energy Economics, 126, 107012.
---> Asks WHY the large-scale models fail so badly in DR (2022) and traces the failure to insufficient carbon sensitivity of sea ice in the large-scale models.

New Penn Master's in Applied Economics and Data Science (MEDS)

I am thrilled to announce that Penn's newly-launched Master's in Applied Economics and Data Science (MEDS) is now accepting applicants for entry in August 2026. 

For information on curriculum, the application process, and faculty (including our fantastic newly-hired Executive Director, Mike Lipsitz), see https://www.lps.upenn.edu/degree-programs/meds. 

Inquiries are already arriving, and there are only thirty spots, but please continue to spread the word to potential students! We firmly believe -- and it seems that students are surely appreciating -- that Penn Economics is uniquely positioned to train leaders for the 21st century at the applied economics + data science interface.

Bobby Mariano Passing

It is with great sadness that I report the passing of Roberto S. ("Bobby") Mariano, Emeritus Professor of Economics at the University of Pennsylvania, on Thursday, April 17, 2025. It was some months ago now, but I just recently re-started this blog, so: 

Bobby received his Ph.D. in Economics from Stanford University in 1970, advised by renowned statistician Theodore W. Anderson. He served on the Penn Economics faculty from 1971 to 2004, teaching a variety of graduate and undergraduate econometrics courses. He then moved to Singapore Management University, where he served as Founding Dean of the School of Economics and Social Sciences (2002-2007) and Dean of the School of Economics (2007-2010). He was well known in econometrics for his pioneering methodological contributions to exact finite-sample distribution theory and predictive accuracy comparison, and he was an elected Fellow of the Econometric Society. In 2022, he was appointed by Pope Francis to serve as a Board Member of the Vatican's Supervisory and Financial Information Authority.  He was an energetic and gregarious colleague, and his bright smile, positivity, warmth, and sense of humor will be dearly missed.

Best conference location in North America: 9th Conference on Econometric Models of Climate Change (EMCC)

 

9th Conference on Econometric Models of Climate Change (EMCC)

Great conference, great community, and yes, great location.  I try to begin all recent climate talks by mentioning this conference and showing this picture, which provides fine motivation for students to start doing climate econometrics...

On tyrants and recalcitrant statistical agencies II: The case of the BLS

[Oh geez, I forgot that the Consumer Price Index (CPI) was coming out today. I had planned to specialize the last post to the BLS eventually -- as regards not just the Jobs Report but also the CPI. So here goes, quickly...]

Obviously the previous post applies to the current Bureau of Labor Statistics (BLS) Jobs Report situation. Let's hope we don't go too far through the playbook. 

But there's more with the BLS situation. BLS produces not only the Jobs Report, but also the CPI, and hence the key data on inflation faced by U.S. consumers. Just as there's a strong temptation for a tyrant to declare weak jobs data as rigged, so too is there a strong temptation to declare increased inflation data as rigged -- and widespread tariffs may produce increased inflation.

So the BLS may be under immense pressure on multiple fronts in the coming months. I fear that we may go significantly through the playbook, particularly as the CPI just happens to be the price gauge to which Social Security benefits are indexed. If a tyrant who wanted to eliminate the U.S. Social Security system could discredit the CPI, then they could discredit inflation-protected Social Security benefits, and then...

On tyrants and recalcitrant statistical agencies

[This is not about the BLS. Really. At least not yet. In a future post I might specialize it to the BLS.]

What's a tyrant to do when they don't get the data they want from Statistical Agency X? Torture Statistical Agency X, of course. 

Here's my sketch of a playbook:

Deflect blame by insisting that the data were "rigged" by politically–motivated actors.

Install a political loyalist to produce "better" numbers.

If Agency X guardrails constrain the loyalist from producing better numbers, then weaken the Agency X guardrails.

If the weakened Agency X guardrails still don't produce the desired result, then weaken the entire Agency X, for example by slashing its budget.

If the weakened Agency X -- with a loyalist at the helm, weakened guardrails, and a slashed budget – – still fails to produce the desired result, then destroy it, for example by eliminating it and assigning its duties to a more pliable agency (perhaps newly created), or even by eliminating its duties ("Who needs this rigged and boring data anyway? Isn't it just a waste of taxpayer money?")

Real Growth as Assessed by ADS Looks OK So Far

A weak jobs report, and much more importantly the related and absurd firing of the BLS commissioner on August 1, have dominated the recent news. But the firing is a topic for a subsequent blog.

For now let's step back and take a look at the broad U.S. employment/growth picture. Employment gains have evidently slowed in recent months, but job destruction as indicated by initial claims has not risen. Initial claims are a key component of the ADS index of real economic activity, and ADS recently and currently continues to indicate "typical", or "average", growth. So: so far so good, but keep your eye on initial claims in future weeks (updates arrive every Thursday). Fingers crossed.

And Please Take Thirty Seconds for This as Well

You can of course view No Hesitations on a desktop or laptop, at URL https://fxdiebold.blogspot.com.  It also formats nicely for viewing on your phone. But on your phone you don't want to have to go to a browser and type the URL. Instead, you can add a No Hesitations icon on your home screen. Then just tap it and you're in! The steps appear below, for both Andriod and iPhone.

How to Add No Hesitations to Your Home Screen

✅ On Android (using Chrome):

1. Open Chrome and go to https://fxdiebold.blogspot.com

2. Tap the three-dot menu (⋮) in the top-right corner.

3. Tap “Add to Home screen”.

4. Edit the name (optional), then tap Add.

5. You may be asked to confirm—choose Add Automatically or drag to place it.

🎉 Done! You’ll now have an icon on your home screen like an app

---

✅ On iPhone (using Safari):

1. Open Safari and go to https://fxdiebold.blogspot.com

2. Tap the Share icon (square with an arrow).

3. Scroll down and tap “Add to Home Screen”.

4. Edit the name if you want, then tap Add.

🎉 Done! A shortcut will appear on your iPhone’s home screen.

PLEASE Take Thirty Seconds to do This

During my three-year vacation, Google Blogger changed the way it sends email notifications of new blog posts, so PLEASE sign up at right.  You can always cancel if my new blogs are not to your liking!