McKinsey says the trade deficit does not matter

What is good for IBM (oops, Lenovo) is good for America.

OK, that is a cheap shot.

Yesterday’s Wall Street Journal published a column by Jon Hilsenrath that drew on work by Diana Farrell and the McKinsey Global Institute. According to Hilsenrath, McKinsey argues:

"Record US trade deficits are not as threatening as they appear (Editor’s note -- let’s hope not, they sure look big), becuase they are driven in part by increasingly profitable US companies producing in places like China, Mexico and India and shipping their goods and services back to the US. ... Overseas profits account for $2.7 trillion in stock market capitalization."

That market capitalization reflects real earnings from US firms abroad: the Wall Street Journal reports US firms earned $315 billion abroad last year; the BEA’s balance of payments data shows that the US earned $237 billion on its direct investment abroad in 2004, and US firms also have equity stakes in other firms.

So, according to McKinsey:

"Far from reflecting the weakness of the US economy, at least a third of the current account deficit is actually evidence of its strength."

We are hearing variants of that argument a lot right now. The US current account deficit is a sign of strength, not a potential weakness. McKinsey’s Global Institute even argues that we should adopt an "ownership based" approach to measuring the trade deficit. Imported goods produced by US firms abroad should not count.

Goldman does not buy the McKinsey argument. Nor do I.

I agree with some of the policy recommendations in the McKinsey study, including its call to cut the budget deficit and to upgrade US skills. But the "owership" based approach to trade does not work.

The trade balance is NOT a dated concept, and it does not need to be superseded to better account for multinational firms with global supply chains. Balance of payments data accounts for global supply chains quite well, so long as US firms accurately report their overseas income. Imports from US firms with operations abroad are still imports - and if the imported goods also embodies "exported’ US content, that will show up as an export to the country doing the assembly work. Expanding global supply chains will cause overall trade to grow. But a gap between a country’s imports and exports still maters: global supply chains do not let a country import more than it exports forever.

A warning: the remainder of this post is quite long.

Why doesn’t an ownership-based approach to trade work? Global supply chains or not, imports are imports and exports are exports. A good produced by an American firm abroad is still an import. Dollars still have to flow out of the US to pay for a product produced abroad. If the imported good is produced by a foreign firm, the foreign firm gets the profits. If the imported good is produced by an American firm, the US firm gets the profits. But even if US firms get the profits, there is still a net outflow of dollars from the US economy that has to be offset by an inflow of capital from abroad. It is not an accident that (net) financial flows into the US track the trade deficit (and even more so the current account deficit), not the "ownership based" trade deficit.

Let’s go back to the example of IBM and Lenovo.

In the McKinsey approach, taken to its extreme, if IBM set up a factory in China to make PCs, using a chip Intel makes in Singapore, a flat panel screen from Korea, a bunch of components made in Taiwan and then installs some software made mostly in the US Northwest on the machine, and then sells that product in the US, the IBM PC does not count as an import, or show up in the trade deficit (presumably McKinsey also nets out any Microsoft software "services" that are installed abroad, and the Intel "chip design" service embedded in the chip made in Singapore).

Suppose, however, IBM agrees to buy computers made by a Taiwanese owned firm (that does its final assembly in China), and then stamps the IBM label on them. That presumably would be an import in the "ownership" based trade deficit.

And then suppose IBM concludes that there is no money to be made selling computers, and sells its PC business -- including the factory it set up in China in my first example -- to a Chinese owned firm. IBM computers become Lenovo computers. The US is not importing any more computers. But the reported trade deficit surges.

That makes no sense.

In the conventional approach, if something made in the US is sold abroad, it counts as an export, even if the US factory is owned by a foreign firm. If something the US exports -- say a microprocessor -- is combined with other electronic components and is imported back into the US as a computer, that is an import, even if a US firm owns the plant that assembles the final product. In my example, both the microchip export and the computer import show up in the balance of payments, but since the value of what the US imports exceeds the value of what it exports (in this particular example), the balance of trade is negative. That is right.

What about the profits the US firms earns abroad? Doesn’t conventional balance of payments accounting ignore them?

Not at all. The profits on US firms’ foreign activities just show up in the "income" line of the balance of payments, not in the trade line.

Alas, the US doesn’t just have assets abroad. We also owe the rest of the world an fairly impressive sum, and we have to pay something on that too.

Fortunately, even though US liabilities exceed US assets, in 2004, the US earned enough abroad to offset, more or less, the payments we make on our foreign debt. Net investment income is about zero - earnings of $365 billion offset payments of $336 billion. US firms earned $237 billion on their overseas operations, but foreign firms earned a reasonable amount on their US operations ($105 billion) too. And the US had to make payments on all the treasury, agency and corporate bonds held abroad, and pay dividends on foreign holdings of US stock.

Bottom line: we in the US don’t earn enough from the activities of US firms abroad and our other foreign assets to finance our trade deficit. Nor will the US be able to draw on its future overseas earnings to pay for future trade deficits. Why not? Obvious: not only are we a net debtor, but our external debt is growing far faster than our assets (leaving aside valuation gains).

Remember, the US does not save enough to finance private firms’ investment here at home, let alone US firms’ investment abroad. Savings is way too old school for new school America. The US borrows from abroad to invest abroad. To put it differently, if a US firm is investing in China, it is doing so with funds the US, in aggregate, borrows from Europe or Japan. If our earnings abroad exceed the cost of the funds we borrow, we still make money on the transaction -- but not near as much as we would if we could finance our overseas investment with our own savings.

One note: the Bureau of Economic Analysis publishes an "ownership-based" framework for the current account. It adds US firms profits abroad (net revenues) to US exports, and foreign firms profits (net revenues) to US imports. The result is a smaller trade deficit, since US firms earn more abroad than foreign firms earn in the US. In 2003, that adjustment leads the deficit on trade to falls from $498 billion to $378 billion But the overall current account deficit does not change -- since the income line turns very negative (-$193 billion). Net US earnings on direct investment are used to reduce the trade deficit rather than US payments on its external debt. That truely is just accounting.

To quote the the BEA: "the current account balance is the same in both sets of accounts."

McKinsey makes a second argument: exchange rate adjustment won’t reduce the trade deficit; since even if the renminbi and rupiah appreciate a bit, the rest of the world will still be cheap.

According to Farrell’s summary of the McKinsey’s work in the Financial Times a while back:

"In 2004, worldwide foreign direct investment flows topped $600 billion, a record. Most of it went to emerging markets, where labour or land costs are one-tenth of those in the US. Even a significant fall in the value of the dollar is unlikely to affect this."

Let’s ignore for a moment the fact that those investment flows that went toward emerging markets went straight into reserves, and were effectively redirected back into the US treasury and mortgage markets. In a bizarre way, FDI flows to emerging economies are the key to the US treasuries ability to finance itself at low interest rates ....

Instead, let’s suppose McKinsey is right: trade and investment flows are no longer that responsive to even large changes in the exchange rates. One possible conclusion is that the dollar is EXTREMELY overvalued, and will have to fall more than people think to bring the US current account deficit down. However, many seem to be drawing on the McKinsey study to draw the opposite conclusion: if the renminbi appreciates, the US will still import lots of stuff from China, but it will pay more for these imports. And US firms will earn less in China. So why bother?

I think they are understating the impact of exchange rate adjustment. I have laid out my arguments before.

Yes, Chinese workers will earn less than US workers even if the renminbi doubled in value, rising from 8 to 4. Yes, there would still be an incentive to shift production to China. Yes, the US would likely pay more for its current Chinese imports.

So what. The overall effect would still be to lower the global US trade deficit.

For one, the incentive to locate new production in China would be reduced. Productivity is not, one assumes, the same in China as in the US. At some exchange rate, the incentive to shift production to China goes away.

More importantly, Chinese workers will see their dollar earnings rise, and they will import more. The US may not make the things that China wants to buy, but if Brazil is able to sell more to China, Brazil will also be able to buy more from the US.

Finally, the incentive to buy renminbi just because the renminbi is expected to go up in the future would go away after a big enough revaluation. Less capital flowing into China means less Chinese reserve accumulation, less Chinese demand for US treasuries and agencies, higher US interest rates, and, at the end of the day, slower US import growth. Don’t forget the capital market channel.

Firms like Walmart that buy from Chinese firms (and firms that operate their own plants in China) would see the price of the goods they import go up if the renminbi is allowed to appreciate. Anyone investing in China now with the expectation that Chinese wages will stay low and the renminbi won’t be allowed to appreciate is betting that current global imbalances will get bigger, without triggering any countervailing moves. They are taking a big risk. The US current account deficit is not small, unless you engage in creative accounting.

A change in the renminbi would put pressure on the profits of some of McKinsey’s clients. We all talk our book, in various ways. But right now, an undervalued renminbi is putting pressure on the wages of workers in the US tradables sector, even as contributes to US corporate profits and generates financial flows that help interest-sensitive sectors of the US economy. Global rebalancing will have losers as well as winners.