Tuesday, March 14, 2017

The Real Exchange Rate and China's Savings

I attended the Forum for Research in Empirical International Trade (FREIT) in Yogyakarta, Indonesia this past week. Marc Muendler and I were debating the source of China's excess savings, over and above its high level of investment. China's high savings rate could have multiple causes (e.g., lack of a social safety net implies a need for precautionary savings, while countries that are growing fast, such as the Asian Tigers, have a clear tendency to save a lot), and various policies could be at play. However, in this post I merely want to lay out how an undervalued real exchange rate, which can be accomplished with a combination of capital controls and foreign reserve accumulation, could increase the rate of savings.

The short story is that an undervalued RER will increase exports and decrease imports, and this equals the gap between savings and investment by accounting identity.

First, why does a high savings rate matter? A high savings rate (over investment) means that China will be exporting both capital and goods to the US, and that it won't be importing many goods. This is due to accounting identity: Savings (S) - Investment (I) = Exports (X) - Imports (M).

Imagine that, at first, S = I, which implies that X = M, while the RER is fairly valued. Now imagine that China devalues its RER by 40%. This change in relative prices means that exports will surge while imports will fall. This will also imply that savings will be less than investment. What is the logic? Let's have a closer look at accounting identities:

Y = C + I + G + X - M

(Where Y = output, C = consumption, I = Investment, G = Government spending, etc.)

Clearly, if the Yuan is weak, then, on one hand, China will export more, and on the other, it will import less. Thus, output will rise. That's straightforward. Total consumption (of domestically produced goods and of imports) may decrease by a little as imports become more expensive, but this effect is likely to be drowned out by changes in the trade balance. It could decrease by more if China also raises domestic interest rates to counter inflation, something which would cause imports to plunge further. How will savings increase? Savings is income not taxed or spent:

S = (Y - T - C) + (T - G)

Total savings equals private and public savings. Fast Chinese growth will also tend to boost T relative to G, and might also boost Y relative to C. If China undervalues its exchange rate, and also raises its domestic interest rates to contain inflation, C could fall, raising savings. There could also be a cultural explanation as Asian countries tend to have high levels of savings in general, although this doesn't explain why savings is larger than investment, or why savings rates have declined recently in Japan and Taiwan (or why growth pre-dated high savings rates in several of these countries). Trade doesn't show up here, but we can substitute in for Y from the equation above, and get:

S = (( C + I + G + X - M ) - T - C ) + ( T - G )

And now simplify to:

S - I  = X - M

Savings minus investment equals exports minus imports. Exports and imports clearly depend on the real exchange rate, and thus so does the gap of savings over investment. I'm not always the biggest advocate for the use of mathematics in economics, but in some cases it can really help for clarification. The above algebra makes this result seem obvious and intuitive, but your average intelligent guy-in-a-bar wouldn't find a link between the exchange rate and savings intuitive or obvious without looking at the math.

A more controversial assertion is that, because of hysteresis, an undervalued exchange rate yesterday could mean you'll have a larger tradables sector today, and thus more savings. On the other hand, in recent years, with weaker Chinese growth and looser monetary policy, the Chinese government in has been trying to prop up the currency, although this doesn't necessarily mean that the Chinese currency is overvalued from the perspective of its tradable sector. But it does mean that this is a topic deserving of further study.


  1. Much of the reasoning is correct but it suffers from two deep misunderstandings in mainstream economics:
    1. National savings does not originate in private savings but mainly in investments. See http://www.asymptosis.com/note-to-economists-saving-doesnt-create-savings.html and http://www.social-europe.eu/2012/11/saving-does-not-finance-investment/
    2. The money for investments may be saved but often it will be newly created endogenous money, especially in China. Heterodox economists have been trying to point this for the last hundred years with little success but let's try again:

    You want to buy a bigger house and go to bank to get a loan of 100k for the price difference. Your request gets approved and you sign a loan agreement where you recognize your liability to pay interest and amortizations according to a certain plan. Then you give the paper to the bank. For the bank that liability of yours is an asset worth 100k and it appears as such in the balance sheet. Then the bank recognizes its liability to pay 100k to your at any moment you want. They record that liability by crediting your current account by 100k. No money gets moved, there are no cash flows. It's all about mutual recognition of legal liabilities. For the counter-party those are assets. All this is very basic double entry accounting. Anyway the mainstream economists who refuse to look at T accounts are not able to see this endogenous money. This implies that they are not able to see the part of financial system in the economy as it really occurs but try to stick to primitive models of banks as insignificant intermediaries, not tremendous sources or drains (during repayment time) of liquidity.

    1. Interesting point. I'll have to think more about this... I agree that money gets created when banks make loans (the money multiplier). Also, I agree that, in principle, the desire to save can shrink GDP and paradoxically shrink total savings.

      In the buying the bigger house example, I'm not sure I understand why no money gets moved. Don't you take 100k from the bank and pay someone else with it? Or, are you saying if that person takes that money and goes and deposits it?

      In this case, thinking about banks adds complication, but I'm not sure it helps to clarify the situation. (However, perhaps you could prove me wrong!) To me, the role of banks are quite important for understanding the financial crisis, and the subsequent slow recovery, however.

    2. Well, yes, when you buy the bigger house then you pay it and the new endogenous money goes to the seller. If in the same bank then that bank liability moves to another counterparty. If to another bank then the bank must take the money from their current account in a third bank. First bank is no more liable to you and gives up an asset. For them it is the same as if you had taken the money in physical banknotes. For the first bank it's the classical case where bank is just an intermediary. In the other bank new money appears. If some bank is short of liquidity then they issue instruments to other banks. Again there the alternatives are to pay from account in a third bank or to pay by recognizing a liability, which creates endogenous money.

      It's important that you can represent any cash flow as accounting entries. In addition much else happens in accounting: If a company owns a house build tens of years before they may perform an adjustment in assets and liabilities to show the present value. Estimated losses in coming years may be booked as expenses now and company may be declared bankrupt because of that. For national economy the most important thing is endogenous money creation. The huge loan amounts in Mediterranean countries before 2007 did not cause any lack of funds in France or Germany. Repayment of loans diminishes amount of available money. Later on imposing various restrictions on bank loans, in themselves necessary, do restrict speed of recovery to some extent.

      I have T-account examples in http://olliranta.wordpress.com/endogenousmoney/

    3. I tried to find out how savings and investments appear in national accounts. A good guide is https://www.bea.gov/national/pdf/nipa_primer.pdf There you find “Account 6. Domestic Capital Account” that essentially claims the following to be an accounting identity:

      NetSaving + ConsumptionOfFixedCapital + StatisticalDiscrepancy =
      GrossDomesticInvestment + CapitalAccountTransactions + NetLendingOrNetBorrowing

      There NetSaving and GrossDomesticInvestment seem to be based on actual data. NetSaving is Income – Consumption. The whole equation seems funny. NetLendingOrNetBorrowing is a balancing item to make sides match. ConsumptionOfFixedCapital is somehow arbitrary parallel version of capital depreciation. The equation does, of course, indicate whether stock of active capital is decreasing or increasing. The claim that it necessarily must balance has no clear basis. Actually the equation strikingly contradicts the usual accounting identity (X-M) = (S-I) + (T-G). My interpretation of the whole is that in reality new domestic loans are a major source of investment funds. Those funny concepts conceal that fact. Most likely nothing will reconcile against actual bank statistics related to loans to public.

  2. When there is full employment, devaluing the currency will lead to an inflationary spiral unless the working class can be suppressed( the price of imported and exportable goods would go up which in effect is a reduction in the real wage).In that case, the increase in the rate and mass of surplus value( or if you like "saving") will appear as a surplus in the current account.In situations where there is a large pool of peasants, such inflationary spirals can be avoided with an export oriented development strategy. In the latter case( Lewis's model), money wages will not rise and the increase in "saving" would be larger compared to the full employment case. In the case of China the devaluation of the currency against the dollar from 1988 to 1995 was accompanied by a 50 percent rise in the Gini coefficient. My criticism of your piece is its abstractness.What is needed is a political and economic analysis of the role of devaluation in the extraction of surplus value.

    1. Fair point.

      "When there is full employment, devaluing the currency will lead to an inflationary spiral unless the working class can be suppressed( the price of imported and exportable goods would go up which in effect is a reduction in the real wage)."

      Yes, this is more-or-less what happened. China could raise domestic interest rates, suppressing demand and keeping inflation under control. I also understand labor isn't so powerful in post-communist China, so export booms don't need to be so inflationary immediately. But a key part of this was controlling the exchange rate. Fast growth in exports certainly would have caused the currency to appreciate.

  3. This comment has been removed by a blog administrator.