@ericvannostrand.bsky.social & I are in @briefingbook.bsky.social on infrastructure. Last week, @jasonfurman.bsky.social argued that real US infrastructure investment has fallen since the pandemic. Eric & I find however that using BEA’s deflator, real highway spending is up 11% since 2019. 1/11 - ThreadSky

ernietedeschi.bsky.social • 3 days ago

@ericvannostrand.bsky.social & I are in @briefingbook.bsky.social on infrastructure.
Last week, @jasonfurman.bsky.social argued that real US infrastructure investment has fallen since the pandemic.
Eric & I find however that using BEA’s deflator, real highway spending is up 11% since 2019.
1/11

Comments

johnfernald.bsky.social•3 days ago

Which tables/series did you use? This is what I see when I look at the BEA capital tables for real government investment in highways and streets (chained index). It doesn't look like your BEA line...or like Jason's.

doughenwood.bsky.social•3 days ago

That’s all very nice but we already have lots of roads and could use some rail.

lucastosborne.bsky.social•3 days ago

📌

ernietedeschi.bsky.social•3 days ago

There’s no question nominal infrastructure spending rose during the pandemic. Total highway & streets construction spending at all levels of government grew by $39B, or 39%, between 2019 & 2023. Looking at all federal nondefense investment, not just highways, the increase is an even-larger $66B
2/11

ernietedeschi.bsky.social•3 days ago

To estimate inflation-adjusted infrastructure, however, we need a deflator.
We looked at 3. The one Jason uses is the National Highway Construction Cost Index (NHCCI) from the Federal Highway Administration. There’s also one from the US Bureau of Reclamation & the official deflator from BEA.
3/11

ernietedeschi.bsky.social•3 days ago

The NHCCI is estimated using winning construction bids posted on State highway websites, so it should in theory incorporate both materials & labor costs. The BEA & USBR deflators are explicit composites of materials costs (PPIs) & labor costs (ECI).
4/11

slittlehale.bsky.social•3 days ago

Are the PPIs used by BEA Final Demand PPIs, thus incorporating contractor mark-ups on material wholesale prices and on labor or costs? @edzarenski.bsky.social has spilled much ink on the importance of not including cyclical variable construction contractor mark-ups. There's alao a great MLR article.

edzarenski.bsky.social•2 days ago

PPIs do not have a final demand index for Highway, only for Nonres Bldgs.

slittlehale.bsky.social•2 days ago

So the nhcci at least attempts to capture not just input prices but markups

slittlehale.bsky.social•3 days ago

Here is the (corrected) link to the 2023 BLS article describing markups in nonresidential construction through 2022.

https://www.bls.gov/opub/btn/volume-12/nonresidential-building-construction-overhead-and-profit-markups.htm

ernietedeschi.bsky.social•3 days ago

As you can see, the NHCCI is an outlier from other measures by 2023. It’s not entirely clear why. NHCCI could be picking up trends the others aren’t. It’s also possible that NHCCI is underweighting labor costs, e.g. if the bids it’s based on did not reflect 100% of final wages.
5/11

ernietedeschi.bsky.social•3 days ago

The choice of deflator determines the path of real highway spending. Using the official BEA deflator, highway spending rose to its highest real level since 2006. Using NHCCI, it fell to its lowest point in decades.
6/11

ernietedeschi.bsky.social•3 days ago

Is there any outside validation for one or the other? Eric & I looked at a few.
First, real infrastructure more broadly defined—not just highways, but all nondefense investment—is up significantly over the pandemic.
7/11

Comments

Posting Rules

Reply