Julie Zhiyu Fu

made a slide deck for my research in finance class on how to make effective slides. Thought it might be helpful for JMCs as well. Here is the link:

fuzhiyu.me/blogs/slide_...

good luck!

Our take: if you believe investors are heterogeneous, quantity data cannot be ignored.

Paper: priceimpactbound.github.io/PriceImpactB...

Finally, why does the "inelasticity" debate matter?

It's not just a number—it's about whether we can understand asset prices using quantity data.

elastic market → quantities are a sideshow, price data are all you need

inelastic → quantities are CENTRAL for prices

“A figure titled ‘Price Impact Bound’ with x-axis ‘Investor Agreement (ρ)’ from 0 to 1 and y-axis ‘Price Impact (M)’ from 0 to 6. A thick curve declines sharply near ρ≈0, passes near (ρ=0.10, M≈3), (ρ=0.50, M≈1), and approaches zero as ρ→1. Two vertical markers at ρ≈0.10 and ρ≈0.75 define the range labeled ‘Typical agreement in data,’ indicated by a double-headed arrow between them. The area under the curve is shaded to emphasize the bound.”

For avg stocks, the price impact bound ≈ 1 with a medium level of agreement (consistent with data).

To argue for a price impact less than 0.1, you really need investors to agree with each other for more than 99%!

We formalize this intuition mathematically in a simple bound:

Price Impact ≥ (σ_p/σ_q) × √(1/ρ - 1)

where agreement ρ can be loosely understood as avg. corr across investors.

The bound needs few structural assumptions, just like the Hansen–Jagannathan bound for SDF

It has to be that even though they disagree on what the right price should be, their demand is inelastic to price changes.

→ A small portfolio flow requires a large price adjustment to clear

The argument rests on a simple observation: asset prices are volatile, yet portfolio flows are small.

You can easily generate little trading alongside volatile prices with large agreement.

But with large disagreement and volatile prices, why don't investors trade more?

It's almost boring to say investors don't agree with each other.

But "investors' demand is inelastic" excites vivid debates.

A new paper w/ Philippe and Lorenzo:

If you acknowledge investors disagree, you MUST accept inelastic market—that trading has large price impact

Our take: if you believe investors are heterogeneous, quantity data cannot be ignored.

Paper: priceimpactbound.github.io/PriceImpactB...

Finally, why does the "inelasticity" debate matter?

It's not just a number—it's about whether we can understand asset prices using quantity data.

elastic market → quantities are a sideshow, price data are all you need

inelastic → quantities are CENTRAL for prices

“A figure titled ‘Price Impact Bound’ with x-axis ‘Investor Agreement (ρ)’ from 0 to 1 and y-axis ‘Price Impact (M)’ from 0 to 6. A thick curve declines sharply near ρ≈0, passes near (ρ=0.10, M≈3), (ρ=0.50, M≈1), and approaches zero as ρ→1. Two vertical markers at ρ≈0.10 and ρ≈0.75 define the range labeled ‘Typical agreement in data,’ indicated by a double-headed arrow between them. The area under the curve is shaded to emphasize the bound.”

For avg stocks, the price impact bound ≈ 1 with a medium level of agreement (consistent with data).

To argue for a price impact less than 0.1, you really need investors to agree with each other for more than 99%!

We formalize this intuition mathematically in a simple bound:

Price Impact ≥ (σ_p/σ_q) × √(1/ρ - 1)

where agreement ρ can be loosely understood as avg. corr across investors.

The bound needs few structural assumptions, just like the Hansen–Jagannathan bound for SDF

It has to be that even though they disagree on what the right price should be, their demand is inelastic to price changes.

→ A small portfolio flow requires a large price adjustment to clear

The argument rests on a simple observation: asset prices are volatile, yet portfolio flows are small.

You can easily generate little trading alongside volatile prices with large agreement.

But with large disagreement and volatile prices, why don't investors trade more?

This makes the basis trade unwinding explanation really unsatisfying to me...can anyone with more institutional knowledge explain to me why this may also happen due to basis trade unwinding?

Traders all say the 10yr yield hike is due to basis trade unwinding. But unwinding has two legs: selling spot and buying future, so future should appreciate. But what we saw is that Treasury future price dropped as well (upper panel), actually pretty much in sync with the spot price (lower panel)

The Nicest Swamp on the Internet Reddit’s not perfect, but it may be the best platform on a junky web.

Reddit is suddenly having a moment. Why? www.theatlantic.com/magazine/arc...

Sent it to JMCs

7 years ago when I just came to the US I was appalled by the perfunctory attitude people hold towards the lunch. A lunch should be a bowl of hot ramen followed by a refreshing nap.

Now I'm the one bringing a bottle of Soylent to the faculty lounge drinking it under derision.

This is the way 🌝

Most curious about 4 and in particular on how to use Git with different type of co-authors: who I can't force to use Git and who can use it minimally.

Please send comments and happy to discuss more :)
SSRN: papers.ssrn.com/sol3/papers....

4.3/ Finally, the idea that a Treasury sell-off by major foreign investors such as China may have strong yield-increasing effect is quite correct in the first decade of the 2000s but less so in recent years -- foreign investors do not contribute much to yield movements since 2015.

4.2/ U.S. banks and foreign investors become substantially more price inelastic after the GFC, while the Fed has stepped up to the game as a state-contingent liquidity provider in the Treasury market. We discuss factors contributing to these regime shifts in detail in the paper.

4.1/ We find little evidence supporting the conventional wisdom that foreigners are the major driver of "flight-to-Treasuries" during risk-off episodes. The pattern is evident even in the raw data, and with our model we can quantify the respective role of foreign & domestic investors.

4/ We can use the estimated system to decompose yield movements and study who and what factors move yields. Then we focus on several important investor groups and discuss three main findings:

3/ The estimated "macro" multiplier is 1 -- 1% positive demand shocks push down yields on a 10-year note by 10bps. This aligns well with estimates in the literature: larger than "micro" multipliers for govt bonds but smaller than the macro multiplier for stocks.

a woman says power is power in a gif from gifsec.com ALT: a woman says power is power in a gif from gifsec.com

2/ this approach yields an estimator that is 1) flexible (each sector has its own residual supply curve shifts that id's the elasticity); 2) intuitive (it has an IV and a bias-corrected OLS interpretation); and 3) powerful (it’s asymptotically efficient), and…

1/ we estimate an asset demand/supply system that can handle heterogeneous price elasticities and factor loadings across investors and time regimes. For identification, we propose a simple yet powerful approach exploiting granular idiosyncratic shocks (Gabaix-Koijen).

First skeet for the first paper as an AP! In *Anatomy of Treasury Market*, we study the drivers of U.S. Treasury yields over the past two decades using a flexible asset demand & supply system (with two amazing coauthors @haonanzhou.bsky.social and @manavchaudhary.bsky.social). Some highlights:

You are on bluesky! (Pun intended)