For the instructor — how to run it (12 minutes; it closes Meeting 1)
Teaching point. A bond that never defaults can still lose a large share of its value when
rates rise. The loss grows with duration: longer maturity, a smaller coupon and a lower starting yield all
make it bigger. Since 2022, Chinese wealth-management investors take that loss directly.
- Slide 11 bond (preset 1), then Show the arithmetic: 3.82 + 3.65 + 90.53 = 97.99, and
Σ t × worth today / price = 2.885. Switch back to the picture: the price is the sum of the dark
bars, and the gold triangle is where they balance.
- 2-year, then 30-year, both at +1 point: −1.85% against −14.63%. The prediction on slide 12
usually says three times; it is about eight.
- 30-year, no coupon: the triangle moves to year 30, duration 30. Push the change to +5 points:
the straight line predicts a negative price. That is convexity — the rule breaks for long bonds
and big moves.
- US 30-year at 5.32%, then China 30-year at 2.15%, both +1 point: about −13% against
−19%. A lower starting yield makes a long bond more sensitive, which is where Session 4 goes.
- Mishkin: 10% → 20% reproduces the textbook — the 30-year returns −39.7% after a
rise from 10% to 20%. The price falls by about half (−49.8%); the straight line says −94%.
Skip this one if the session is behind.
What students get wrong. They think a government bond cannot lose money; they mix up the coupon
rate and the yield; they treat maturity as the measure of risk; they expect price and yield to move the same
way.
Debrief question. Two banks each hold 100 of government bonds, one in 2-year bonds and one in
30-year bonds (4% coupons, 4.735% yield). Rates rise one point. How much does each lose? (−1.8 and
−12.9 per 100 of face value.) Which bank’s depositors should worry? That is Silicon Valley Bank,
Week 4.