September 18, 2020
I have long enjoyed reading Matthew Klein’s columns in the Financial Times and elsewhere. They are invariably insightful and I have learned much from them. I am less familiar with Michael Pettis’ work, but I have liked what I have read. Therefore, I expected a lot from their book, Trade Wars are Class Wars, and I was not disappointed.
The basic point is that the major trade imbalances in the world over the last four decades have been driven by the suppression of wage growth, with income being redistributed from labor to capital. This has led to shortfalls in aggregate demand that countries try to offset by having trade surpluses. The main actors in that picture are China and Germany.
In the Klein-Pettis view, the U.S. has also suffered from this upward redistribution, although it has taken a somewhat different form since the country has run persistent trade deficits over this period. While I largely agree with this framing, I have some minor quibbles with the story they layout and one very large one.
In the minor quibble category, Klein and Pettis (KP) criticize Trump adviser Peter Navarro for focusing on the bilateral trade deficits the United States runs with China and other countries. I have no stake in defending Peter Navarro, but at least some of us who are concerned about the trade deficit with China have argued that the U.S. should be pressing China to raise the value of its currency relative to the dollar.
If the renminbi rose against the dollar, it should mean, other things equal, that China would have a lower trade surplus and therefore fewer savings, and the United States would have a smaller trade deficit, and therefore more savings. This isn’t an issue of a policy being thwarted by the inescapable logic of accounting identities, it is a question of the direction of causation in these identities. And, if we don’t think the U.S. economy is at full employment, more net exports will mean higher GDP and more savings.
I should also mention here that KP make the valid point that the direct trade balance between China and the U.S. does not accurately measure the net flows between the two countries, since much of the price of items exported from China is due to inputs from other countries. Therefore, our imports from China overstate the value of goods produced in China and then exported to the United States.
While this is true, it neglects the flip side, that many of the items we buy from Japan, the European Union, and elsewhere, include inputs from China. If we want to pull out the third country value-added from the goods imported from China, to get a better measure of bilateral trade flows, we also have to add in the Chinese value-added in the items imported from third countries. While KP are not guilty of the one-sided adjustment, other economists are. It is one of the items in the Games Economists Play textbook.
Another point that is at least under-emphasized in KP, is that real wages soared in China in the period since 2000, in spite of the wage repression measures they describe. Real wages have risen at close to double-digit annual rates over this period. While they would have risen even faster if the country had not maintained an undervalued currency, and other measures to suppress wage growth, it has to be easier politically to maintain a policy of repressing wage growth in a context where wages are rising rapidly than when they are stagnating or falling.
KP also hold out the prospect that China is going to run up against some limit as it encourages companies to pursue wasteful investment projects that leave them ever further in debt. It’s not clear why a country that issues its own currency and has massive foreign reserves and a trade surplus, would ever hit a limit. I suppose if somehow excessive debt leads to explosive demand growth, and an inflationary spiral, it would be a serious problem. But it is hard to see how we get from where we are now, with persistent shortfalls in demand as the main issue, to the complete opposite problem.
My last quibble on the surplus country side is the claim that the assets issued or guaranteed by the U.S. government are risk-free. I know this is the standard wisdom, but I have no idea what it is supposed to mean.
I get that the U.S. government is not about to default on its debt. That can be treated as a near certainty. But the dollar can and often does fall in value relative to other currencies. In fact, since the start of 2020, the dollar has fallen by more than 6.0 percent against the euro. I’m not sure why the prospect that the dollar could fall in value relative to other currencies should not be seen as a risk. I don’t think most people would consider an investment that loses 6.0 percent of its value in eight months as risk-free.
On the U.S. side, KP for some reason ignore the wealth effect when discussing the low savings rates in the United States in the 1990s and the 2000s pre-crash. The idea of a wealth effect on consumption is not new, nor to my knowledge especially controversial.
The 1990s stock market bubble created close to $10 billion in bubble wealth. Assuming a wealth effect of 3-4 percent, this implied increased consumption on the order of 3-4 percent of GDP. That fits well with the drop in savings we saw over this period. As a practical matter, this manifested itself in smaller contributions to 401(k)s and also declining employer contributions to defined benefit pensions. The rise in the stock market was effectively making the contribution for employers. Some of us warned about the problem this was creating for pensions at the time, but the irrational exuberance of the day drowned us out.
There was a similar story with the housing bubble. As KP note, people were happy to borrow against the rapidly growing equity in their homes. With the bubble creating roughly $8 trillion in additional housing equity, and a housing wealth effect of 4-6 cents on the dollar, the run-up in house prices can fully explain the fall in saving rates in the bubble years.
My reason for harping on this issue is that many economists were surprised when consumption plunged in 2008-2010. There was no basis for this surprise. The bubble that had been driving consumption had been deflated. Savings rates returned to roughly their long-term average, and have stayed there even as the debt overhang from the crash dissipated. There is nothing surprising in this story that needs explaining.
Okay, but enough of the quibbles, my big objection is the failure to focus adequately on the structuring of trade. This is a very central part of the wage suppression story, at least in the United States.
First, it is important to make the point that in the United States the upward redistribution was for the most part not from wages to profits, but from the wages of ordinary workers to the wages of highly paid workers like CEOs and other top executives, Wall Street and Silicon Valley types, and highly paid professionals.
KP for some reason see a very different picture of the data than me. They report a 7.0 percentage point drop in the labor share of national income from 2000 to 2007 (p 179). I see a drop of 1.2 percentage points (NIPA Table 1.10, Line 2 divided by the sum of Line 2 and Line 9).
Furthermore, even this drop in labor shares is questionable. Profits were inflated during these years by the profits earned in the financial sector from issuing dubious loans that subsequently went bad. This would be similar to companies reporting profits on non-existent sales. There were not real profits and of course, the banks lost hugely when these loans went bad in 2008—2010.
There was a drop in labor shares in the weak labor market resulting from the Great Recession. This was being reversed with the recovery, but we still had some way to go when the pandemic hit. By my calculations, just over 10 percent of the gap between productivity growth and median wage growth from 1979 to 2019 can be explained by this shift from labor to capital. The rest is due to upward redistribution within the wage structure.
This gets us to the structure of the trade. Our trade policy was focused on reducing or eliminating barriers to merchandise trade. This has the predicted and actual result of lowering the wages of manufacturing workers, which spilled over to putting downward pressure on the wages of non-college-educated workers more generally. When workers in the United States had to compete with workers in China and elsewhere, earning less than one-tenth of their pay, it was inevitable they would either take large pay cuts or lose their jobs.
However, trade liberalization did not have to focus on manufactured goods or at least exclusively focus on manufactured goods. We could have focused trade policy on reducing barriers that prevent qualified professionals from working in the United States. Doctors and dentists in the United States earn roughly twice as much as their counterparts in other wealthy countries. Trade liberalization in these professions could have brought their pay closer to the OECD average, saving households more than $100 billion annually from their healthcare bill. We didn’t see the liberalization of licensing restrictions (they were actually made tighter in the 1990s) because doctors and dentists have far more political power than autoworkers.
The other part of this story that is almost completely absent from KP is the tightening of rules on patent and copyright monopolies, which have been a key part of every trade deal for the last three decades. These monopolies are forms of protectionism, the polar opposite of free trade. They lead to prices for prescription drugs, medical equipment, software, and other protected items that are often hundreds of times the free market price, the equivalent of tariffs of tens of thousands percent.
They also are associated with enormous upward redistribution, as ordinary workers have to pay much higher prices for the benefit of the people with the skills associated with producing these products. Bill Gates would likely still be working for a living if Microsoft had to sell software without government-granted copyright or patent monopolies.
This is a huge issue going forward, as the cost of protection for intellectual products is likely to continue to grow as a share of the economy. It is already huge. We will spend over $500 billion this year on prescription drugs alone. If these drugs were sold in a free market, we would almost certainly pay less than $100 billion. The difference of $400 billion is more than $3,000 per family. This $400 billion in potential savings is almost 80 percent of what we will spend on cars and parts in 2020.
In future trade deals, the terms of protection for intellectual products, both foreign and domestic, will swamp the importance of trade in manufactured goods. The loss of good-paying jobs in manufacturing was a huge deal in the 1980s, 1990s, and especially the 2000s, but unfortunately, we don’t stand much to lose anymore.
Manufacturing employment is now less than 9.0 percent of total employment. Furthermore, while manufacturing jobs used to provide relatively good-paying employment for workers without college degrees, this is no longer true. The average wage for production and non-supervisory workers in manufacturing is 6.0 percent less than for production and non-supervisory workers in the private sector as a whole. This gap would be reduced when factoring in benefits, but the idea that a manufacturing job is a good-paying job where a worker could support a family and have health care and a pension is simply no longer true.
The reason for the deterioration in the quality of manufacturing jobs is not a secret. The unionization rate in manufacturing has plummeted, largely due to trade. Until the pandemic hit in March, we had added back more than 1.6 million manufacturing jobs from the Great Recession trough in 2010. Nonetheless, the number of union members in manufacturing had fallen by almost 900,000. In short, as we added back jobs in the sector, they were overwhelmingly lower paying non-union jobs. There is little reason to believe that these workers were much better off than workers in an Amazon distribution center or other service sector jobs.
This points to the fact that manufacturing jobs should not be at the center of the trade agenda for those concerned about reversing the upward redistribution of income. The focus should be on reducing the protectionist barriers that raise the pay of the most highly paid professionals, and on limiting the protections for intellectual property that raise the prices we pay for a wide range of items and hand the money to a relatively small elite in a position to benefit.
As in real war, when it comes to trade, it is important that we not be fighting the last war. We have to go beyond KP to be properly prepared for battle.
 The corresponding overstatement on the output side would be treating the issuance of the loan as adding value. In reality, a bad loan is not adding value to the economy.