One of the reasons why trade-related discussions can seem so off-the-mark, I think, is because the conditions governing international trade are much more complex than we often realize. The determinants of the international balance of trade basically include anything that affects domestic consumption and domestic production, which pretty much means nearly everything in economics. Among other things this means that there is a very wide range of government policies that can affect trade – sometimes explicitly and sometimes implicitly.
Unfortunately much of the analysis and debate doesn’t seem to get this. For example, many economists have pointed out that the bailout of GM is effectively a protectionist measure. I think it clearly has a trade impact, and this impact is “protectionist”, although not intended that way. What is missing from the discussion, I think, is a clear explanation of why it is effectively a protectionist measure. I would argue that the GM bailout has a trade impact because it affects in specific ways the balance between production and consumption in the US (and, of course, elsewhere), and since the US trade deficit is also the gap between US consumption and US production, to the extent that the bailout affects this gap it must affect the US trade balance.
In that case we can posit at least two obvious ways in which the bailout affects the gap. First, by effectively subsidizing the cost of producing GM cars, it increases automobile production in the US. Second, by allowing GM to retain the workers it would have otherwise fired, it increases consumption in the US by the amount which the retained GM workers spend on consumption, i.e. some large fraction (depending on their savings rate) of their wages. At first I was going to suggest that the relevant number was actually not their wages but the difference between their wages and their welfare payments, since most of the workers would presumably continue to earn some money after they were fired, but then it occurred to me that their welfare payments would have reduced other government spending, so perhaps it is not relevant (this is a subject of much disagreement between Keynesians and monetarists).
Since the GM bailout almost certainly increases production by more than it increases consumption, its direct impact is to reduce the US trade deficit (although to be complete we would also need to consider how the GM bailout affects GM’s competitors, many of whom produce cars in the US). However there are of course secondary impacts, the most important of which is the funding of the bailout. If funding the money used to bail out GM ended up crowding out investment in other production facilities, then the question becomes whether or not those other production facilities would have involved a more productive use of the money and, therefore, had a better impact, either in the short term or in the long-term on total US production. This is also part of the debate between Keynesians and monetarists.
So because we typically think of the currency and tariff policies as the main tools to affect trade – which usually means to boost net exports – much of the discussion surrounding trade policies tends to be limited to these two issues (although when the subject of “dumping” comes up the discussion becomes a lot more sophisticated). This leads to strange arguments. For example when I talk about an increase in trade frictions leading to an increase in trade protection, I am often countered by the argument that the WTO makes tariffs very difficult so that protection becomes almost impossible. This is manifestly not true, but more on that later. These, at any rate, are the two best-known trade-related policies:
» Currency policies, whose first-order impact is to determine the relative pricing of imports and exports, but there are also a series of second-order impacts that can be very important.
» Tariffs, especially import tariffs, whose first-order impact also determines the relative pricing of traded goods, usually imports.
To repeat, any policy that affects the relationship between production and consumption must affect the trade balance because the excess of production over consumption is the trade surplus (or deficit, of course, if consumption exceeds production). So currency policies affect the trade balance primarily by their impacts on diverting production and consumption. A country, for example, that devalues its currency, raises the cost of imported goods and so reduces the real value of wages. This of course usually causes total consumption to decline. At the same time it allows local producers who compete with imports, who might not have been as cost effective as foreigners at the previous exchange rate, to begin producing more goods for sale to the domestic market. The combination of reducing domestic consumption somewhat and increasing domestic production means that the trade deficit will decline or the trade surplus increase.
One thing that economists always point out, and contrary to the mercantilist view, is that this increases domestic unemployment, but it doesn’t necessarily improve local welfare. Remember that by devaluing its currency, a country is worsening the terms of trade for its own products – it must now produce more stuff locally for export to pay for the same amount of imports. It also results in a net reduction in total consumption. Currency policies often involve a tradeoff between employment and total welfare in the short term. Over the long term it is not always clear that this is true, however.
Trade tariffs work in very much the same way. Devaluing the currency by 10%, for example, has the same impact as putting a 10% subsidy, or negative tariff, on exports (the government pays exporters an amount equal to 10% of the value of their exports) and a 9.1% tariff on imports.
But, as I hope these examples show, it is not just tariff and currency policies that affect the trade balance. Anything that affects the gap between consumption and production also affects the balance of trade. These include:
» Corporate and personal income taxes. Personal income taxes reduce consumption by reducing disposable income. Corporate income taxes reduce production by raising costs for producers.
» Sales and other taxes. Depending on their impact they can also affect trade. The most obvious case is a sales tax which, by raising the cost of goods, reduces real wages and so reduces consumption.
The impact of taxes on trade are complicated by the fact that taxes represent a transfer of resources, so to understand fully their impact we also need to know what the government does with new tax revenues or how it finances reduced tax revenues. These can enhance or reduce either consumption or production.
There are a lot of other factors that impact trade, and I have randomly included the following, which I think are especially important, at least in China. Others can and will have others to add or may dispute some of my arguments.
» The level of worker’s wages. They impact trade in two ways – by affecting consumption via affecting the purchasing power of households, and by affecting production via the cost to businesses of labor, and they tend to work in the same way as far as the trade balance is concerned. Lowering wages reduces consumption and increase production, so as to have a positive impact on the trade balance. Needless to say many economists have pointed out that low wages in China are one of the reasons for the high trade surplus.
» Unemployment benefits. Unemployment benefits tend to cause consumption to decline more slowly than production when factories close for obvious reasons, although of course we need to take into account how these benefits are funded. I would guess that when a country’s workers do not receive unemployment benefits, it tends to be “positive” for the trade.
» Subsidized costs to producers – electricity, oil, commodities, etc. Subsidizing the cost of production is a very effective way to increase exports since it directly increases production by increasing the returns to producers. It also has an impact, albeit much smaller, on increasing consumption via its impact on employment. Since these subsidies are financed by taxes, subsidies may also constrain consumption somewhat, depending on the nature of the taxes used to fund subsidies.
» Subsidized costs to consumers. This boosts consumption, of course, although with the same caveat as above – its net effect depends on how the subsidies are financed.
» Corporate lending rates. This should be included in “subsidized costs to producers” but I put it in a separate category because it is a very important type of subsidy, especially in China. Low interest rates for manufacturers of course make it much easier to borrow money to fund otherwise unprofitable production facilities, thereby increasing production (and increasing consumption somewhat by its impact on employment). If the lending is directed at non-manufacturing activities, such as to the service sector, it will not spur manufacturing production but will still increase consumption. As an aside, in my May 20 entry I discuss an HKMA study that argues that in China 100% of SOE profitability can be explained by interest subsidies which, I argue, actually understate the true value of those subsidies, suggesting that many SOEs would actually be value destroyers if it were not for subsidized financing. This is clearly a very important reason for the Chinese trade surplus, in my opinion.
» Deposit rates. In last week’s entry this claim generated a certain amount of controversy in the comments section, but it is widely believed that in some countries, like China, reducing deposit rates causes savings to increase and consumption to decline. I discuss some possible reasons in my November 27 entry, the most important of which is probably that in high savings countries in which most savings are in the form of bank deposits, the interest earned on banking deposits is a significant fraction of total disposable income, and lowering deposit rates has an effect similar to lowering wages. Of course if this is true, low deposit rates are likely to reduce consumption, just as low lending rates to producers are likely to increase production. They may also increase production by reducing the opportunity cost for corporations of investing retained earnings.
» Other credit intervention – lending guarantees, directed lending, forbearance on addressing NPLs, etc. This is fairly complex since there are many ways to intervene in credit, but any policy which increases the provision of credit to manufacturers must increase production directly. It increases consumption somewhat too, as in the two previous cases, by creating employment and thus raising the total amount of wages paid. If the lending policy increases credit provision to consumers or the non-manufacturing sector, it increases consumption directly. Credit for infrastructure investment is a little more complex since it probably increases consumption today and production tomorrow.
» Special mention: cleaning up NPLs. This really belongs in the category above but in the case of China it deserves its own entry. There are two ways to recapitalize banks suffering from a surge in NPLs. One way is to recapitalize them directly. When the government does this is simply transfers money to the banks, as China did before the IPOs of the major banks. Depending on how these transfers are funded, they can have a variety of effects on production and consumption. The second way is to guarantee banking profitability by keeping a wide spread between lending and deposit rates. Policymakers may also keep lending rates very low in order to slow the accumulation of NPLs and make it easier for marginal borrowers to survive. As I discuss above, this can result in very low deposit rates, which constrains consumption, and very low lending rates, which increases production.
I focus a lot on various financial sector issues because it seems to me that it is through the banking system that policymakers can have their largest impact on the trade balance. By keeping rates excessively low (and remember that almost all interest rates in China are either fully controlled and set by the PBoC or very heavily affected by the controlled interest rates), policymakers can boost production and constrain consumption quite easily.
When production grows faster than consumption this necessarily forces an increase in the savings rates – which ties this entry into my previous entry. And of course by controlling the direction of credit, either directly or indirectly by implicit or explicit government guarantees, the government has a major say in the total amount of production. It is probably not a coincidence that in the countries that followed export-oriented growth policies, the so called Asian development model, interest rates and credit tended to be highly controlled either directly or indirectly by the government and regulators. These countries have all had “surprisingly” low interest rates and banking systems that channeled funding mostly into the manufacturing sector. I would argue that a controlled banking sector is a very important tool for trade policy.
Another one of the issues that this opens up is the distinction I have made many times between total consumption and net consumption. Notice that many policies increase both production and consumption. They usually do the latter by increasing employment. In many cases Chinese policies have been successful in boosting consumption in just this way. Since the world manifestly needs more consumption, as US household consumption declines precipitously, anything that boosts Chinese consumption should be a good thing, right?
Maybe not. What the world needs from China is not an increase in total consumption but rather an increase in net consumption – i.e. the excess of new consumption over new production – that is roughly in line with the decline in US net consumption. If consumption grows, but production grows just as fast, or even faster (and we can tell by looking at the trade balance corrected for various pricing effects and one-off purchases or sales), then the world imbalance is getting worse and the overcapacity problem will not have been addressed.
This means that many policies that may seem on the surface to be purely domestic policies are actually trade policies too, and legitimately subject to scrutiny and even criticism from abroad. This is clear from the GM bailout. I don’t believe that Congressmen agreed to the bailout because they wanted to engage in protectionist behavior. They did so because they wanted to protect American jobs, but they did so in a way that almost inevitably has a trade impact. The same thing is happening in China, but there is a real reluctance to consider that policies aimed, for example, at limiting unemployment among aluminum plant workers in Hunan (or is it Henan?) are not just internal matters but also international trade policies.