SJTU 2026 Macroeconomics

Student presentations

Presentation topics

Your presentation is a small piece of real research. You pick a question, you answer it with one of the models we covered in the lectures, you bring real data, and you finish with your own interpretation. The seven topics below are ready to use: each one names the course tool it builds on, the data you can download for free, and the figures you could build.

How to build a good presentation

  • Aim for about 12 to 15 minutes. That is enough for one question, one model, three or four charts and a clear answer.
  • One chart carries one message. If you need two sentences to explain what a chart shows, split it into two charts.
  • Label both axes and write the data source under every chart. A chart without a source is a rumour.
  • Use a ratio (log) scale for anything that grows over decades. On a normal scale, steady growth looks like an explosion; on a ratio scale, steady growth is a straight line.
  • End with a one-sentence answer to your question. If you cannot say your conclusion in one sentence, you have not finished the analysis yet.
  • If you present as a group, practise the handover between speakers. A clean handover is the easiest thing to prepare and the most common thing to go wrong.

What I look for

When I watch your presentation, I am looking for five things:

  1. A clear question. One sentence, stated at the start, answered at the end.
  2. Correct use of the course tool. Growth accounting, the Solow diagram, the Phillips curve, AS/AD, IS-MP, the (r − g) snowball or the saving-investment identity, used the way we used it in the lectures.
  3. Honest data work. Real series, correct units, sources cited, and no chart that hides what the data actually show.
  4. An interpretation that follows from the evidence. Tell me what your charts let you conclude, and also what they do not.
  5. Clear delivery. Plain language, a steady pace, and charts the back row can read.

Timing, group size and scheduling will be announced in class.

01 The Chinese Growth Miracle: Capital, Labour or Ideas? Lectures 3–4 · Jones 4–6

Why this matters

China's growth since 1980 is the largest and fastest rise in living standards ever recorded for so many people. Growth accounting lets you split that growth into three measurable parts: more capital, more (and better educated) labour, and higher productivity (TFP). The split matters, because the Solow model says growth from capital deepening must slow down, while growth from ideas can continue.

Where it sits in the course

Builds on Lecture 3 (Jones 4–5): growth accounting with the production model and the Solow diagram, and Lecture 4 (Jones 6): growth from ideas.

Questions to answer

  1. (Theory) Starting from Y = A K1/3L2/3, write down the growth-accounting equation. Which parts are measured directly, and why is TFP called "the residual"?
  2. (Data) Using the Penn World Table, compute the contributions of capital per worker, human capital and TFP to China's growth, decade by decade, from 1980 to 2019. Which part did the most work in each decade?
  3. (Data) Plot China's GDP per capita as a percentage of the US level, at PPP, from 1980 to today. Is the gap still closing at the same speed?
  4. (Theory) What does the Solow model predict for growth that is driven mainly by capital deepening? Where on the transition path does your evidence place China today?
  5. (Interpretation) Given your TFP numbers and what Lecture 4 says about ideas, what would have to be true for fast growth to continue at higher income levels?

Data you can use

  • Penn World Table 10.01 via rug.nl/ggdc (variables: rgdpna, rnna, emp, hc, labsh, rtfpna).
  • World Bank WDI: NY.GDP.PCAP.PP.KD (GDP per capita, PPP, constant international $) and NY.GDP.MKTP.KD.ZG (GDP growth, annual %).
  • NBS, data.stats.gov.cn: annual national accounts, employment and investment tables.
  • Maddison Project Database via rug.nl/ggdc: long-run real GDP per capita for context before 1980.

Figures you could build

  1. Log real GDP per capita, China and the US, 1980 to 2024, ratio scale: two lines, and the vertical gap between them narrows, then narrows more slowly.
  2. Stacked bars by decade: the share of China's growth explained by capital deepening, human capital and TFP. The reader should see which bar shrinks over time.
  3. China's GDP per capita as a percentage of the US level (PPP), one line, 1980 to 2024: catching up, with the slope your evidence for whether convergence is slowing.

A suggested plan

  1. Hook with one number: an economy growing at 9 per cent a year doubles roughly every 8 years (the Rule of 70), and China grew near that pace for about three decades.
  2. Set up the tool: the production model and the growth-accounting split from Lecture 3.
  3. Show your data: the decade-by-decade accounting table and your three figures.
  4. Interpret: how much was capital, how much labour and education, how much TFP, and what Solow says about each part's future.
  5. Verdict in one sentence: your answer to "capital, labour or ideas?", stated plainly.

One common pitfall

Comparing incomes across countries at market exchange rates: for level comparisons you need PPP, exactly as we discussed in Lecture 2.

02 Ageing and Growth: China's Demographic Transition Lectures 3 & 5 · Jones 4–7

Why this matters

China's working-age population has already peaked and is now falling, and the fertility rate is close to one birth per woman, among the lowest in the world. Labour is one of the three inputs in our production model, so this is a first-order question for long-run growth. Japan, Korea and Germany went through this transition earlier, which gives you real evidence to work with.

Where it sits in the course

Builds on Lecture 3 (Jones 4–5): labour in the production model and the Solow diagram, and Lecture 5 (Jones 7): the labour market.

Questions to answer

  1. (Theory) In Y = A K1/3L2/3, suppose L falls steadily while A and the saving rate are unchanged. What happens to total output, and what happens to output per worker? Keep the two answers separate.
  2. (Data) Plot the working-age share of the population for China, Japan, Korea and Germany from 1960 to 2024, and extend China with the UN projections. When did each country peak?
  3. (Data) Compare fertility rates and the share of the population aged 65 and above across the same four countries. How fast is China ageing compared with Japan at the same stage?
  4. (Interpretation) What actually happened to Japan's and Germany's growth as they aged? Compare growth of GDP per person with growth of GDP per working-age adult.
  5. (Interpretation) Which offsets look strongest in the data: more education (human capital), more capital and automation, or higher participation and later retirement?

Data you can use

  • UN World Population Prospects 2024 via population.un.org: working-age population 15–64, estimates and projections, China / Japan / Korea / Germany.
  • World Bank WDI: SP.POP.1564.TO.ZS (population ages 15–64, % of total), SP.POP.65UP.TO.ZS (population ages 65+, % of total), SP.DYN.TFRT.IN (fertility rate, births per woman).
  • OECD Data Explorer, data-explorer.oecd.org: labour-force participation rates by age group.
  • NBS, data.stats.gov.cn: population and employment tables.

Figures you could build

  1. Working-age share of the population, four countries, 1960 to 2050 (projection dashed): a set of humps peaking at different dates, China's peak clearly behind it.
  2. For Japan: growth of GDP per person against growth of GDP per working-age adult, by decade: the per-worker bars stay respectable even when the headline slows.
  3. Share of the population aged 65 and above, China against Japan with Japan shifted about 20 years earlier: the two lines nearly on top of each other.

A suggested plan

  1. Hook with one number: the UN projects China's working-age population to fall by well over 100 million people between now and 2050.
  2. Set up the tool: labour in the production model, and what the Solow diagram says when an input shrinks.
  3. Show your data: the four-country hump chart and your Japan comparison.
  4. Interpret: how much of the ageing effect can education, capital and participation realistically offset, using the earlier agers as evidence.
  5. Verdict in one sentence: does ageing mean slower growth in total GDP, in GDP per person, or both?

One common pitfall

Mixing up total GDP and GDP per person: ageing can slow the first a lot while the second keeps growing, and a presentation that blurs the two proves neither.

03 The 2021–2023 Inflation Surge in the Euro Area and the UK Lectures 6, 8–9 · Jones 8, 11–14

Why this matters

In October 2022, UK inflation reached 11.1 per cent, the highest in over forty years, and euro-area inflation peaked above 10 per cent in the same month. Two years later both were back near 2 per cent, and neither economy went through a deep recession. This episode is a live test of everything in the second half of the course: supply shocks, the Phillips curve and the role of anchored expectations.

Where it sits in the course

Builds on Lecture 6 (Jones 8): inflation and the Fisher equation, Lecture 8 (Jones 11–12): the Phillips curve, and Lecture 9 (Jones 13–14): AS/AD.

Questions to answer

  1. (Data) Compute 12-month inflation from the euro-area HICP and the UK CPI index. When exactly did each peak, how high, and how fast did each come down?
  2. (Theory) In AS/AD, a supply shock and a demand shock leave different fingerprints on output and prices. Which pattern fits 2021–2023, and what role did energy prices play?
  3. (Data) Plot the ECB's policy rate over the episode, and describe the Bank of England's Bank Rate path alongside it. Who moved first, and how far did each go?
  4. (Interpretation) Unemployment barely rose while inflation fell. What does the Phillips curve with anchored expectations say about how that is possible, and what would un-anchored expectations have looked like instead?
  5. (Theory) Using the sacrifice-ratio idea from Lecture 8, why did this disinflation cost so much less than the Volcker disinflation of the early 1980s?

Data you can use

  • FRED: CP0000EZ19M086NEST (euro-area HICP, all items, index) and GBRCPIALLMINMEI (UK CPI, all items, index): compute the 12-month percentage change yourself.
  • FRED: ECBDFR (ECB deposit facility rate, daily, from 1999).
  • Bank of England, bankofengland.co.uk: the official Bank Rate history.
  • Eurostat: HICP tables and the monthly unemployment rate (table une_rt_m); ONS, ons.gov.uk, for UK CPI releases and labour-market data.

Figures you could build

  1. Twelve-month inflation, euro area and the UK, 2019 to 2025, with the two peaks marked: a sharp mountain, up fast and down almost as fast.
  2. Policy rates as step lines over the same window: rates near zero until 2022, then the steepest climb in decades.
  3. A Phillips-curve scatter for the euro area, unemployment against inflation, monthly 2019 to 2024, points joined in time order: a tall loop, not a nice downward curve.

A suggested plan

  1. Hook with one number: 11.1 per cent, UK inflation in October 2022.
  2. Set up the tool: the AS/AD fingerprints of supply against demand shocks, plus the Phillips curve with expectations.
  3. Show your data: the inflation mountain, the policy-rate steps and the Phillips loop.
  4. Interpret: how much was energy, how much demand, and why anchored expectations made the descent cheap.
  5. Verdict in one sentence: why inflation fell without a deep recession, in your own words.

One common pitfall

Plotting the CPI index level and calling it inflation: inflation is the 12-month percentage change, and the two charts tell completely different stories.

04 Japan at the Zero Lower Bound: Lessons from Three Decades Lectures 6 & 9 · Jones 8, 13–14

Why this matters

Between the mid 1990s and the early 2010s, Japan's consumer price level barely moved, and its policy interest rate sat at or near zero for longer than most of you have been alive. Japan hit the zero lower bound a full decade before the US and Europe did, so it is the world's longest-running natural experiment on deflation, the ZLB and quantitative easing.

Where it sits in the course

Builds on Lecture 9 (Jones 13–14): AS/AD, the ZLB, deflation and quantitative easing, and Lecture 6 (Jones 8): the Fisher equation and the costs of deflation.

Questions to answer

  1. (Data) Compute Japan's 12-month inflation from the CPI index. Which years show outright deflation, and how large was it?
  2. (Theory) Use the Fisher equation to explain the deflationary spiral: when prices fall and the nominal rate cannot go below zero, what happens to the real interest rate, and why does that deepen the slump in AS/AD?
  3. (Data) Plot Japan's long-term government bond yield from 1989 onward. What does its long slide toward zero tell you about expected inflation and expected growth?
  4. (Interpretation) What changed after 2013, when the Bank of Japan adopted large-scale easing and a 2 per cent inflation target, and what does the return to positive interest rates in 2024 suggest about whether the exit finally arrived?
  5. (Interpretation) Which of Japan's lessons did the US and the euro area actually apply after 2008, and did acting faster produce a better outcome?

Data you can use

  • FRED: JPNCPIALLMINMEI (Japan CPI, all items, index, monthly, 1955 to 2021); for the most recent years use the Statistics Bureau of Japan CPI tables at e-stat.go.jp.
  • FRED: IRLTLT01JPM156N (Japan long-term, 10-year, government bond yield, monthly, from 1989) and JPNRGDPEXP (Japan real GDP, quarterly, from 1994).
  • FRED: QJPN628BIS (Japan real residential property prices, BIS series) for the 1990 bust that started it all.
  • IMF World Economic Outlook database and OECD Data Explorer: annual output gaps, inflation and per-capita growth for Japan against peers; Bank of Japan, boj.or.jp, for policy dates.

Figures you could build

  1. Japan's 12-month CPI inflation, 1980 to today, with a horizontal line at zero: long stretches below the line, then the recent climb back above it.
  2. The 10-year bond yield, 1989 to today, one line sliding from around 8 per cent to essentially zero and only recently lifting: three decades of low expected inflation and low expected growth in a single line.
  3. Log real GDP per person, Japan and the US, on a ratio scale: the honest comparison, and it makes Japan look much less like a disaster than the headline numbers do.

A suggested plan

  1. Hook with one number: Japan's 10-year bond yield spent years near 0 per cent, something textbooks once said should not happen.
  2. Set up the tool: the ZLB and the deflationary spiral in AS/AD, with the Fisher equation doing the work.
  3. Show your data: the inflation chart with its zero line, the bond-yield slide and the per-person growth comparison.
  4. Interpret: what conventional policy could not do at the ZLB, what QE did and did not achieve, and what 2013 to 2024 changed.
  5. Verdict in one sentence: the single most important lesson other central banks should take from Japan.

One common pitfall

Calling the whole period "lost decades" without checking per-person figures: Japan's growth per person held up far better than total GDP, because the population was shrinking.

05 One Currency, Many Governments: Greece and the Euro-Area Debt Crisis Lecture 10 · Jones 18

Why this matters

In early 2012, Greece's 10-year government bond yield rose above 25 per cent while Germany, using the same currency, borrowed for close to 2 per cent. Greece then went through the largest sovereign debt restructuring in history, and its real GDP fell by about a quarter. The episode is the clearest real-world demonstration of the debt arithmetic we built in Lecture 10.

Where it sits in the course

Builds on Lecture 10 (Jones 18): the government budget constraint, the (r − g) snowball and self-fulfilling debt crises.

Questions to answer

  1. (Theory) Write down the law of motion for the debt-to-GDP ratio and explain the (r − g) snowball: why do a rising interest rate and a shrinking economy feed the ratio at the same time?
  2. (Data) Plot the spread between Greek and German 10-year yields from 2001 to 2015. When did markets start treating the two borrowers differently, and how violently?
  3. (Data) Trace Greece's debt-to-GDP ratio and its real GDP through the crisis. How much of the ratio's rise came from new borrowing, and how much from the collapsing denominator?
  4. (Interpretation) Why can a government that borrows in its own currency not be forced into default the way Greece was, and what does that say about who the euro's rules really constrain?
  5. (Interpretation) In July 2012 the ECB promised to do "whatever it takes", and yields fell before it bought a single bond. What does that tell you about self-fulfilling debt crises?

Data you can use

  • FRED: IRLTLT01GRM156N (Greece long-term, 10-year, government bond yield, monthly) and IRLTLT01DEM156N (Germany, same series): the spread is one subtraction.
  • IMF World Economic Outlook database: general government gross debt (% of GDP), real GDP growth and the primary balance for Greece.
  • Eurostat: government finance statistics and quarterly national accounts for Greece and the euro area.
  • FRED: CLVMNACSCAB1GQEA19 (euro-area real GDP, quarterly) as the benchmark path to compare Greece against.

Figures you could build

  1. Greek and German 10-year yields, 2001 to 2015, two lines: glued together for a decade, then torn apart after 2009.
  2. Greece's debt-to-GDP ratio with the crisis years shaded: the ratio climbs fastest exactly when GDP is falling, the snowball in one picture.
  3. Real GDP indexed to 100 in 2007, Greece against the euro area: the euro area dips and recovers, Greece falls about 25 points and stays down for years.

A suggested plan

  1. Hook with one number: over 25 per cent, the yield Greece faced in 2012 in the same currency Germany borrowed in for about 2 per cent.
  2. Set up the tool: the budget constraint and the (r − g) snowball from Lecture 10.
  3. Show your data: the spread chart, the debt ratio and the GDP comparison.
  4. Interpret: separate the arithmetic (r, g and the primary balance) from the institutions (who prints the currency, and what the ECB's promise changed).
  5. Verdict in one sentence: what made Greek debt unmanageable when larger debts elsewhere were not.

One common pitfall

Reading the rise in debt-to-GDP as pure over-borrowing: in the crisis years the falling denominator, GDP itself, did much of the work, and austerity that shrinks GDP can raise the ratio it is meant to cut.

06 Housing Booms and Busts: How Property Moves the Macroeconomy Lectures 7 & 9 · Jones 9–10, 13–14

Why this matters

US house prices fell by roughly a quarter between 2006 and 2012, and the fall set off the worst global financial crisis in eighty years. Yet Japan after 1990 and Spain and Ireland after 2008 show that similar price falls play out very differently depending on leverage and balance sheets, and China's recent property adjustment gives you a fourth, still-unfolding comparison. The interesting question is not whether prices fall, but when a fall becomes a financial crisis.

Where it sits in the course

Builds on Lecture 7 (Jones 9–10): the anatomy of a financial crisis, leverage and balance sheets, and Lecture 9 (Jones 13–14): the risk-premium wedge in AS/AD.

Questions to answer

  1. (Data) Using the BIS real house-price indices, align four episodes so that each market's peak equals 100 at time zero: the US from 2006, Japan from 1991, Spain or Ireland from 2007, and China from 2021. How do the four descents compare in speed and depth?
  2. (Theory) Using the leverage arithmetic from Lecture 7, show how a 25 per cent fall in house prices can wipe out the equity of a highly leveraged household or bank, and why that turns a price fall into forced selling.
  3. (Data) What happened to unemployment and output in each episode? US unemployment reached 10 per cent in 2009; how do the other cases compare?
  4. (Interpretation) Why did the US bust become a global banking crisis within months, while Japan's and China's adjustments have been slower and more contained? Consider who held the debt, and how much of the economy construction made up.
  5. (Interpretation) From your four cases, what conditions turn a housing bust into a financial crisis, and which of them are policy choices?

Data you can use

  • BIS residential property price statistics, bis.org/statistics: real house-price indices for all four economies; several are mirrored on FRED as QUSR628BIS (US), QJPN628BIS (Japan), QESR628BIS (Spain), QIER628BIS (Ireland) and QCNR628BIS (China).
  • FRED: CSUSHPINSA (Case-Shiller US national home price index) and UNRATE (US unemployment rate).
  • OECD Data Explorer and Eurostat: unemployment and construction shares for Japan, Spain and Ireland.
  • NBS, data.stats.gov.cn: floor space and real-estate investment tables for the Chinese case.

Figures you could build

  1. Real house prices, four episodes aligned with each peak = 100 at time zero, quarters since peak on the x-axis: four descents at very different speeds on one chart.
  2. Two small panels for the US: Case-Shiller falling in one, unemployment rising to 10 per cent in the other, same time axis, so the reader sees prices lead the labour market.
  3. Construction plus real-estate share of GDP at each episode's peak, one bar per country: the taller the bar, the harder the landing that followed.

A suggested plan

  1. Hook with one number: about 25 per cent, the fall in real US house prices that was enough to freeze the world's financial system.
  2. Set up the tool: leverage and balance sheets from Lecture 7, plus the risk-premium wedge from Lecture 9.
  3. Show your data: the aligned four-episode chart and your US panels.
  4. Interpret: rank the four episodes by damage, and explain the ranking with leverage, who held the debt, and construction's weight in the economy.
  5. Verdict in one sentence: your answer to "when does a bust become a crisis?".

One common pitfall

Using nominal house prices: over long episodes inflation hides much of the story, so use the BIS real (inflation-adjusted) indices.

07 Why Do Nations Save So Differently? China's Saving Rate and Global Imbalances Lectures 2 & 10 · Jones 2, 18

Why this matters

China saves more than 40 per cent of its GDP, one of the highest rates ever recorded for a large economy; the UK saves less than half that share. The saving-investment identity says every one of those saved yuan must become either investment at home or lending abroad, so national saving rates quietly shape investment, current accounts and the whole pattern of global capital flows. This topic is pure accounting done well, and the accounting is more powerful than most opinions.

Where it sits in the course

Builds on Lecture 2 (Jones 2): the identity Y = C + I + G + NX and the saving-investment identity, and Lecture 10 (Jones 18): global imbalances.

Questions to answer

  1. (Theory) Starting from Y = C + I + G + NX, derive S = I + CA, national saving equals investment plus the current account. What must be true, as a matter of accounting, when a country saves more than it invests?
  2. (Data) Plot gross saving as a share of GDP for China, the US, Germany, Japan and the UK from 1980 to today. How unusual is China, and has its rate started to drift?
  3. (Data) For China, plot saving, investment and the current account together. Does S = I + CA hold in the data, and in which periods was the gap between saving and investment largest?
  4. (Interpretation) Which explanations for China's high saving does the evidence support: demographics and the age structure, precautionary saving by households for health, education and retirement, or high corporate saving? What kind of data would separate them?
  5. (Interpretation) Low-saving economies show the mirror image: investment above saving and a current-account deficit. What does the identity tell you about such deficits, and, just as important, what does it not tell you about their causes?

Data you can use

  • World Bank WDI: NY.GNS.ICTR.ZS (gross savings, % of GDP), NE.GDI.TOTL.ZS (gross capital formation, % of GDP), BN.CAB.XOKA.GD.ZS (current account balance, % of GDP).
  • IMF World Economic Outlook database: total investment and gross national savings (% of GDP), with forecasts.
  • NBS, data.stats.gov.cn: flow-of-funds and household survey tables for the household / corporate / government split of saving.
  • UN World Population Prospects 2024: age structure, if you pursue the demographic explanation.

Figures you could build

  1. Gross saving as a share of GDP, five countries, 1980 to today: China's line sits far above the pack for three decades.
  2. For China: saving and investment as two lines with the gap between them shaded, and the shaded gap is the current account, the identity as one picture.
  3. A cross-country scatter for a recent year: gross saving on one axis, gross capital formation on the other, one dot per country, with the 45-degree line drawn in; distance from the line is the current account.

A suggested plan

  1. Hook with one number: more than 40 per cent of everything China produces in a year is saved, roughly triple the UK's share.
  2. Set up the tool: derive S = I + CA on one slide, from the Lecture 2 identity.
  3. Show your data: the five-country saving chart and your China identity picture.
  4. Interpret: weigh the demographic, precautionary and corporate explanations against your evidence, and show the mirror image in a low-saving economy.
  5. Verdict in one sentence: the main reason nations save so differently, as your evidence sees it.

One common pitfall

Treating the current account as a policy lever someone sets: it is the accounting difference between saving and investment, so any explanation of it must go through S and I, not around them.

Your own topic. You may also propose a topic that is not on this list. The rule is simple: it must apply one of the course's tools to real data. Send me your question and your planned data source, and get my approval before you start work.
Shanghai · China Course home Teaching CV
University of Oxford