Opening 0:00
Welcome to today's China report. Today: the macro-financial bind. China's enormous trade surplus, a currency a lot of people think is held down on purpose, and a debt pile somebody eventually has to eat. Four pieces published this summer, four answers to the same question, who pays. Carnegie, the Council on Foreign Relations, the Peterson Institute, and the Jamestown Foundation. One at a time.
Carnegie on who paid for the debt cleanup 0:32
Michael Pettis, for the Carnegie Endowment, twenty-fifth of August, in their China Financial Markets series.
His claim: China's famous bank cleanup at the end of the nineteen nineties didn't get rid of any losses. It moved them. Four asset management companies bought about one point four trillion renminbi of bad loans, roughly twenty percent of GDP, at full face value, funded by bonds sold straight back to the banks they were cleaning. Circular. So the loss got paid somewhere else.
Here's the mechanism. For a decade after, Beijing held deposit rates below inflation and far below nominal growth. Savers, overwhelmingly households, earned less than nothing. Borrowers, state firms, local governments, banks, pocketed the difference. Pettis puts that transfer at as much as five points of GDP a year. And his headline number: household consumption fell from sixty-three point nine percent of GDP in two thousand to forty-nine point four percent by twenty ten.
What's it built on? Official bad-loan estimates of twenty-five to thirty percent of the big four banks' books, which he rejects, in favour of the World Bank's forty to fifty. Asset management mechanics from the Yale Journal of Financial Crises. The five-point figure from IMF research.
And that's my problem. His opening move is that Chinese official statistics understated the losses by half. Fine. But that fourteen-point collapse in the consumption share comes out of the same statistical system, and he takes it straight. You can't be a sceptic about the bank numbers and a literalist about the national accounts in the same essay. An independent reconstruction of the household income share would settle it, and he doesn't offer one.
One outside number. Korea's government-funded Korea Institute for International Economic Policy published a brief in April putting China's return on invested capital at around eleven percent through the two thousands, down to six now. Eleven percent. The investment households were subsidising was earning something. Pettis would say the subsidy is what made it look profitable, fair enough, except he never looks at returns at all.
His forecast has no date on it. Just a dilemma: the household option is used up, so the next bill splits between business and government, out in the open. No clock, so nothing to catch him out with.
CFR on the undervalued renminbi 3:26
Brad Setser, Council on Foreign Relations, second of August, in Follow the Money.
The renminbi is undervalued by about twenty percent on a straight calculation, thirty to thirty-five once you adjust for gold imports and misreported investment income. Held there on purpose: the central bank sets the daily midpoint, and state banks buy something like two billion dollars a trading day to stop it rising. A one point four trillion dollar goods surplus, he says, isn't coming down on domestic reform alone.
The chain is tight. The twenty twenty-two depreciation, ten to fifteen percent against the dollar, was followed on standard lags by an export boom, net exports added about one and a half points to growth in twenty twenty-four, and again in twenty twenty-five. So nominal moves do produce real ones. That's aimed at Gopinath, Gourinchas and Rey, who say appreciation just gets eaten by deflation. Setser says look at the data: deflation has moderated, not accelerated.
The evidence is mostly the Fund's, its exchange rate model, its twenty twenty-five staff report. The two-billion-a-day figure isn't published intervention data. It's inferred from banking flows, because the whole point of routing purchases through state banks is that reserves don't move. He's open about that.
Here's what bothers me. His fifth point is that the IMF's own forecasts of a shrinking Chinese surplus have not been borne out. He's right, and it's damning. But the twenty percent undervaluation comes out of that same framework, take a current account gap of about three points of GDP, divide by an elasticity. If the model can't predict the current account, the gap it computes isn't independent evidence of anything. He leans on the Fund where it agrees with him and beats it where it doesn't.
His forecasts do have clocks, to his credit. A seventh cumulative point of growth from net exports by next year. Car exports possibly twenty million a year within three years, roughly one in three cars sold outside China. I went looking for an independent projection of that from the Japanese and Korean institutes that track Asian auto supply chains, and I couldn't find one.
Peterson Institute on the property bust 5:59
Tianlei Huang at the Peterson Institute, seventh of August. The shortest piece here.
The surplus is a housing story, not a subsidy story. Compare twenty ten to twenty twenty against twenty twenty-one to twenty twenty-four. Household gross capital formation, mostly buying homes, fell four point one points of GDP, thirteen percent down to eight point nine. Household saving barely moved: twenty-one point nine to twenty-one point eight. So the household surplus widened by nearly four points, and it left the country as net exports. Remedy: fiscal expansion aimed at transfers and the safety net. Tariffs treat the symptom.
The mechanism is a sectoral balances identity, saving minus investment, across households, government and firms, equals the current account. And that's the weakness, because an identity holds by construction. It can't tell you which side moved first.
Now look at his own numbers. The government balance deteriorated four point eight points, more than the household balance widened. On his own chart, government dissaving more than absorbs the household swing. So if the current account still went from roughly nothing to two percent of GDP, the move has to be coming from the corporate sector. Which is precisely where the subsidy story lives. His own decomposition points at the sector the piece was written to rule out.
It's Chinese official flow-of-funds data plus the Fund's staff report. One chart, no independent series. And there's no dated forecast in it at all, just a hope that equity markets might soak up household saving if firms stay optimistic. That's not a prediction.
Jamestown on outbound capital controls 8:03
Christopher Nye and Charles Sun for the Jamestown Foundation, eleventh of July.
State Council Order Number Eight Three Seven took effect on the first of July. For the first time, China's outbound investment rules cover resident individuals, not just companies. Approval, national security review, and real penalties: forced disposal of assets, confiscation, bans on investing abroad of one to three years. Article thirty-three reaches personal money in overseas markets. Add May's regulatory action against Tiger Brokers, Futu and Longbridge, and a two-year wind-down where mainland accounts can only sell and send money out. After twenty twenty-eight, mainland-facing platforms shut down entirely.
The reasoning: households found a grey route out through offshore brokers, regulators hit the brokers, now they're hitting the investors, and there's no grandfather clause, so existing offshore holdings sit legally exposed. Conclusion: the machinery keeping household savings at home is being tightened, not dismantled.
This is document analysis, and it's careful. Articles quoted, the regulator's notice, the implementation plan. But notice what's missing. The one number telling you how big this channel was is Futu's own disclosure, client assets up nearly sixty-six percent, to over one point two trillion Hong Kong dollars, about a hundred and sixty billion US, by the end of twenty twenty-five. And Futu attributes that growth chiefly to Hong Kong and Malaysian clients. Not mainland ones. So the report's only quantitative evidence doesn't measure the leak it's describing. Everything else is legal capability, not observed flow, and the authors concede it, saying whether this reaches ordinary private wealth depends on implementing rules nobody has published yet.
Their forecast is the sharpest of the four, because it has a date: offshore platforms serving the mainland gone by twenty twenty-eight.
What to watch 10:23
Nearest thing to resolve is Setser's, whether net exports add another point and a half to growth by the end of this year, or whether the surplus finally rolls over. Then twenty twenty-eight, for whether that broker shutdown actually lands.
And there's a real disagreement buried underneath. Setser says the central government's net debt is close to zero, so it has room to spend, and Huang's entire remedy depends on that being true. Pettis says the fiscal option is the politically hard one precisely because the debt is already the problem. They can't all be right, and which one is wrong decides who pays.