Governments manipulate currencies by saving

There are two views of current account surpluses.  One is that they reflect “undervalued” currencies.  Another is that they reflect saving/investment imbalances.  Thus the CA surplus is the capital account deficit, which is (by definition) domestic saving minus domestic investment.

As long as the term ‘undervalued’ is purely descriptive, with no normative implications, then both views are correct.  David Glasner has a very good post on the Chinese CA surplus and China’s exchange rate policy.  I mostly agree, but will take up his challenge in the following quotation.  I think I can offer a perspective that clarifies some of the issues David raises.

That leaves another, less focused, method by which governments can offer protection from international competition to certain industries or groups. The method is precisely for the government and the monetary authority to do what Keynes’s Law says can’t be done:  to choose an exchange rate that undervalues the currency, thereby giving an extra advantage or profit cushion to all producers of tradable products (i.e., export industries and import-competing industries), perhaps spreading the benefits of protection more widely than governments, if their choices were not restricted by international agreements, would wish. However, to prevent the resulting inflow of foreign cash from driving up domestic prices and eliminating any competitive advantage, the monetary authority must sterilize the induced cash inflows by selling assets to mop up the domestic currency just issued in exchange for the foreign cash directed toward domestic exporters.  .  .  .  But to borrow a concept from Austrian Business Cycle Theory, this may not be a sustainable long-run policy for a central bank, because maintaining the undervalued exchange rate would require the central bank to keep accumulating foreign-exchange reserves indefinitely, while selling off domestic assets to prevent the domestic money supply from increasing. The central bank might even run out of domestic assets with which to mop up the currency created to absorb the inflow of foreign cash. But in a rapidly expanding economy (like China’s), the demand for currency may be growing so rapidly that the domestic currency created in exchange for the inflow of foreign currency can be absorbed by the public without creating any significant upward pressure on prices necessitating a sell-off of domestic assets to prevent an outbreak of domestic inflation.

It is thus the growth in, and the changing composition of, the balance sheet of China’s central bank rather than the value of the dollar/yuan exchange rate that tells us whether the Chinese are engaging in currency manipulation. To get some perspective on how the balance sheet of Chinese central banks has been changing, consider that Chinese nominal GDP in 2009 was about 2.5 times as large as it was in 2003 while Chinese holdings of foreign exchange reserves in 2009 were more than 5 times greater than those holdings were in 2003. This means that the rate of growth (about 25% a year) in foreign-exchange reserves held by the Chinese central bank between 2003 and 2009 was more than twice as great as the rate of growth in Chinese nominal GDP over the same period. Over that period, the share of the total assets of the Chinese central bank represented by foreign exchange has grown from 48% in December 2003 to almost 80% in December 2010. Those changes are certainly consistent with the practice of currency manipulation.  However, except for 2009, there was no year since 2000 in which the holdings of domestic assets by the Chinese central bank actually fell, suggesting that there has been very little actual sterilization undertaken by the Chinese central bank.  If there has indeed been no (or almost no) actual sterilization by the Chinese central bank, then, despite my long-standing suspicions about what the Chinese have been doing, I cannot conclude that the Chinese have been engaging in currency manipulation. But perhaps one needs to look more closely at the details of how the balance sheet of the Chinese central bank has been changing over time.  I would welcome the thoughts of others on how to interpret evidence of how the balance sheet of the Chinese central bank has been changing.

I’d like to point out that while the central bank often does the job of boosting domestic saving (and hence depreciating the real exchange rate) by purchasing lots of foreign assets, there is no logical reason to connect this “saving” policy with monetary policy.  The Chinese government could have another agency (the Treasury?) tell exporters to sell their dollar earnings to them.  This agency would raise yuan funds via taxes (not printing money) and then use these yuan funds to buy dollars from Chinese exporters.  The agency would then presumably swap the dollars for interest earning dollar assets like T-bonds.  This would put downward pressure on the Chinese real exchange rate for the exact same reason that sterilized Chinese central bank purchases of dollars would put downward pressure on the exchange rate.

Actual Chinese exchange rate manipulation usually involves three factors:

1.  More Chinese government saving.

2.  The saving is done by the central bank.

3.  The central bank keeps the nominal exchange rate pegged.

But only the first is important.  If the Chinese government saves a huge percentage of GDP, and total Chinese saving rises above total Chinese investment, then by definition China has a CA surplus.  And this surplus would occur even if the exchange rate were floating, and if the purchases were done by the Chinese Treasury, not its central bank. That’s why you often see huge CA surpluses in countries that don’t have pegged exchange rates (Switzerland (prior to the recent peg), Singapore, Norway, etc).  They have government policies which involve either enormous government saving (Singapore and Norway) or policies that encourage private saving (Switzerland.)  It should also be noted that government saving does not automatically produce a CA surplus. Australia is a notable counterexample.  The Aussie government does some saving, but the private sector engages in massive borrowing from the rest of the world, so they still end up with a large CA deficit.

Why does the Chinese currency manipulation seem “worse” than other countries?  Obviously China is big, so it gets noticed.  And obviously China often does peg its nominal exchange rate.  That draws the attention of the 99.99% of policymakers who have never read Mankiw’s intro textbook, and don’t realize that government saving (not pegged exchange rates) causes CA surpluses.  If the PBoC pegged the rate but didn’t accumulate foreign assets, then they’d be forced to inflate, and the real exchange rate would rise.  Or (as David points out) if they bought foreign assets but didn’t sterilize the purchases then they’d produce inflation, which would push the real exchange rate higher.

In addition to the currency peg, China also has its central bank do lots of saving by accumulating massive holdings of foreign assets.  That confuses even those people who do know that the real issue is the government saving, not the fixed exchange rate.  They forget that it makes no difference whether the extra government saving is done via central banks, or via things like sovereign wealth funds, tax breaks on saving, etc.  So China gets noticed.

In the end none of this should matter, as the job situation in the US is determined by two factors:

1.  US supply-side policies

2.  US NGDP growth (i.e. monetary policy.)

China plays no role in either, unless we are stupid enough to let it play a role in #2.  If so, we have no one but ourselves to blame.

I don’t know enough about the PBoC balance sheet to comment on David’s specific claim.  My hunch is that you’d want to look at the entire balance sheet of the Chinese government.  If the total saving of the Chinese government seems “high” relative to some arbitrary standard, you could call that currency manipulation. I have absolutely no idea, however, how much the Chinese government should be saving.  So while the Chinese government is almost certainly trying to manipulate the real exchange rate, I have no idea whether the real exchange rate would be all that different if there was no intentional “manipulation,” just Chinese saving done for the usual reasons.  China’s real exchange rate is not particularly unusual for a low income country.

Most countries save too little, as government policies are usually biased toward current consumption, and double-tax patient people who prefer future consumption.

Memo to liberals:  If the world saved as much as it should, there would almost certainly be a massive boom in “infrastructure” investment.  A sound monetary policy would also help.


Tags:

 
 
 

23 Responses to “Governments manipulate currencies by saving”

  1. Gravatar of Justin Irving Justin Irving
    4. November 2012 at 09:00

    If I understand this view of current account/capital account, a resource boom has no systematic impact on the current account. If the government responds like Norway, monopolizing the resource and directing much of it into foreign asset purchases, obviously the current account swells. But the resource boom only ‘helps’ the CA if it leads to relatively more saving.

    So would it be safe to assume that, as U.S. oil production rises in the next decade (seems likely), the current account will basically be unaffected? Americans will simply import more non-oil goods and services?

  2. Gravatar of Jon Jon
    4. November 2012 at 10:00

    The PBoC creates domestic assets by issuing bonds in the name of PBoC. This s how they have pegged the real rate…

    In this way they will never run out of domestic assets to sell, although the process is constrained by the interest payments on those bonds. When the Chinese shifted the term structure of their asset base–which is 90% foreign–they suddenly had an issue sterilizing due to reduced coupons. then they tried jacking up the reserve ratios to contain inflation. This worked moderately but at high cost… And so in turn they allowed some appreciation in recent years; they really didn’t have a choice.

  3. Gravatar of Costard Costard
    4. November 2012 at 11:36

    “China plays no role in either, unless we are stupid enough to let it play a role in #2. If so, we have no one but ourselves to blame.”

    You’re suggesting an arms race. China can increase dollar asset purchases just as easily as we can ease monetary policy, and these two actions will not offset. Forex manipulation favors specific industries, and monetary easing benefits different ones. You might maintain employment this way, but not without distributional effects that will bite you in the +++ the moment either country throws in the towel. A higher U.S. inflation rate would tend to discourage this sort of game, but twenty years in, with the imbalances in both countries severe, you’d simply be playing with fire. A Chinese rebalancing will have an effect on employment, but it will have even greater effect on the credibility of the dollar.

    I’ll say this: it would have been better to use monetary policy to float employment than to use fiscal (and tax) policy like we did. Issuing new debt allowed the net exporters to buy dollar assets without the inflationary and currency effects that these purchases would otherwise have had. And the $16 trillion outstanding guarantees that, should the trend reverse, our national finances will go with it.

  4. Gravatar of Saturos Saturos
    4. November 2012 at 12:05

    Great post. In fact Mankiw has a bunch of great old blog posts making similar points:
    http://gregmankiw.blogspot.com.au/2006/03/chinese-exchange-rate.html
    http://gregmankiw.blogspot.com.au/2007/05/turnabout.html

    Scott, did you see the Bartlett tweet I posted on the Greeley-page?

  5. Gravatar of Saturos Saturos
    4. November 2012 at 12:06

    Here is probably an even more blunt post: http://gregmankiw.blogspot.com.au/2006/05/mechanics-of-fixed-exchange-rate.html

    And here is Michael Spence: http://gregmankiw.blogspot.com.au/2007/01/spence-on-trade-deficit.html

    and Joseph Stiglitz (makes a good point for a change): http://gregmankiw.blogspot.com.au/2006/05/im-with-joe.html

  6. Gravatar of Jim Crow Jim Crow
    4. November 2012 at 14:43

    Quick question. Even though the FED affects employment with monetary policy, is it the case that ASSUMING no change in the behavior of US fiscal policy that China’s exchange rate pegging has a measurable affect on the structure of US jobs in the tradable sector? For what its worth, I’ve always assumed China’s overt subsidies and heavily controlled banking system had a lot more to do with their export performance than the peg but the peg is just easier for Congress to wrap its head around along with the annual kabuki currency manipulator theater.

  7. Gravatar of ssumner ssumner
    4. November 2012 at 15:18

    Justin, No, I don’t think that would be a safe assumption.

    Saturos, Yes, Bruce Bartlett told me that story back in June. I didn’t mention it in the blog because he said it was off the record.

    Jim, Yes, but there’s no reason why we should care about the distributional effects, unless we also think free trade is bad.

  8. Gravatar of Major_Freedom Major_Freedom
    4. November 2012 at 15:23

    ssumner:

    In the end none of this should matter, as the job situation in the US is determined by two factors:

    1.  US supply-side policies

    2.  US NGDP growth (i.e. monetary policy.)

    2 is incorrect.

    NDGP does not finance wages, period. In fact, in the aggregate, every form of spending is in competition with every other form of spending. When you spend money on final output, you are not paying wages. If everyone only spent money on final output and nothing on labor, then wage paying jobs would disappear, and yet NGDP would remain positive, and would tend to be a function of the money supply (quantity theory).

    You are doing workers no good by buying output.

    As John Stuart Mill remarked a few years ago, “demand for commodities is not demand for labor.”.

  9. Gravatar of gmacd gmacd
    4. November 2012 at 17:20

    Major_Freedom, if you spend money on final output, wouldn’t the producer be the one paying wages? Labor is a cost that would be factored into the price of final goods, and so wages paid must be dependent upon the nominal value of final goods being sold.

    Therefore, it would seem that if output is being bought, then producers would have money to pay their costs, one of which being labor. So workers actually would be benefiting by output being purchased.

  10. Gravatar of Major_Freedom Major_Freedom
    4. November 2012 at 18:18

    gmacd:

    Major_Freedom, if you spend money on final output, wouldn’t the producer be the one paying wages? Labor is a cost that would be factored into the price of final goods, and so wages paid must be dependent upon the nominal value of final goods being sold.

    My point is that when one is spending money on consumer goods, one is not spending money on labor. Sumner’s argument was that NGDP determines employment. One however cannot assume that spending money on final output implies, creates, or presupposes the existence, current or prior, of money spent on labor. They are different expenditures.

    It is possible for a positive NGDP to exist without a single nickel paid to wages.

    Therefore, it would seem that if output is being bought, then producers would have money to pay their costs, one of which being labor. So workers actually would be benefiting by output being purchased.

    Suppose I went into my backyard and cut a tree down, and widdled a baseball bat, and then sold it to you for $10. Final output exists, spending and hence a price for final output exists, and yet no wage payments were made.

    Yes, there are wage payments in our economy and yes there is an NGDP, but wages are not financed out of NGDP. NGDP and wage payments can move in opposite directions.

    The tacit assumption being made by Sumner is the Keynesian doctrine that as long as aggregate demand grows at a particular rate, or does not fall below a particular rate, that somehow wage payments will take care of themselves. This doctrine is theoretically fallacious, and for those non-economists who need historical data to be convinced of economic principles, this doctrine was empirically falsified during the 1970s when price inflation (and NGDP) rose, and yet employment did not, indeed it fell.

  11. Gravatar of gmacd gmacd
    4. November 2012 at 18:49

    Major_Freedom:

    “My point is that when one is spending money on consumer goods, one is not spending money on labor.”

    So now we’re talking about consumer goods, not all goods. I totally agree with you that if John buys a good from Phil, then John is not spending money on labor. But Phil is, because part of the costs associated with the making of that good is labor, so wages are being paid.

    “One however cannot assume that spending money on final output implies, creates, or presupposes the existence, current or prior, of money spent on labor. They are different expenditures.”

    The only examples I can think of where the purchase of a good doesn’t imply that wages are being paid are slavery or a futuristic society where all output is produced by robots.

    “Suppose I went into my backyard and cut a tree down, and widdled a baseball bat, and then sold it to you for $10. Final output exists, spending and hence a price for final output exists, and yet no wage payments were made.”

    In that example you’ve paid yourself the wage.

    “NGDP and wage payments can move in opposite directions. […] this doctrine was empirically falsified during the 1970s when price inflation (and NGDP) rose, and yet employment did not, indeed it fell.”

    So you’re saying that unemployment rose during this period while NGDP rose as a result of inflation (implying that real growth was flat/negative).

    So let’s take the example of a year where NGDP increases 1%, RGDP decreases 1% and inflation increases 2%. Unemployment also increases, so there are less people working. NGDP can be defined as the sum total of incomes of individuals living in a country during the year. Income can come from wages, corporate profits, investment income, farmers’ income, and nonincorporated business income. If wages fall but NGDP rises, then one of those other four sources of income must have risen to compensate.

    So I’d agree that the two can go in opposite directions. The question then is if NGDP were targeted at 5% then would wages grow too? I would say they would in nominal terms since if business owners knew that the Fed was targeting 5% NGDP growth then they could just withhold raises during a downturn instead of laying people off. Real wages would fall but people would keep their jobs in the short run.

  12. Gravatar of Costard Costard
    4. November 2012 at 19:59

    “Jim, Yes, but there’s no reason why we should care about the distributional effects, unless we also think free trade is bad.”

    And yet the distributional effect in the U.S. is to undervalue labor and overvalue capital, causing unemployment — something you do care about.

    How is monetary policy going to resolve this? The labor market will want easing while capital markets need tightening, and accommodating either one will have repercussions for the other. Would you prefer unemployment or asset bubbles?

    More generally, if AD is not the problem, then how can monetary policy be the appropriate solution?

  13. Gravatar of ssumner ssumner
    5. November 2012 at 04:55

    Costard, Trade doesn’t raise the unemployment rate. Stable monetary policy (steady NGDP growth) doesn’t cause asset price bubbles.

  14. Gravatar of About » Blog Archive » Ignoring China Is Not A Great Idea About » Blog Archive » Ignoring China Is Not A Great Idea
    5. November 2012 at 10:40

    […] have to respond to Scott Sumner’s recent post on China. After reviewing China’s exchange-rate manipulation/saving/export policy, Sumner […]

  15. Gravatar of Costard Costard
    5. November 2012 at 10:42

    “Costard, Trade doesn’t raise the unemployment rate.”

    Neither does NGDP. I thought it was clear we were discussing shifts in equilibrium, not equilibrium per se. If China’s trade policies cause manufacturing to contract, there will be a spike in unemployment; and if those same policies also result in increased investment, NGDP won’t necessarily fall. In essence the natural rate of unemployment has risen.

    You say that the job situation is determined by NGDP growth, but the fact remains that the NGDP growth required to maintain a specified unemployment rate can and does change. Monetary policy will not resolve trade distortions, because these result in an imbalance between capital and labor while money is (supposedly) neutral. If you’re watching unemployment, you’re only getting half the picture.

    “Stable monetary policy (steady NGDP growth) doesn’t cause asset price bubbles.”

    This is precisely what I’m disputing. If the relationship between wages and capital is changing, you can have asset price bubbles and steady NGDP growth at the same time. All that is required is a fall in labor’s share of GDP. The Fed didn’t understand this, and they ignored the housing bubble because NGDP and unemployment seemed to justify a low rate policy. They were wrong, and it was because they didn’t understand China.

  16. Gravatar of On the Manipulation of Currencies « Uneasy Money On the Manipulation of Currencies « Uneasy Money
    5. November 2012 at 18:56

    […] time of the asset and liability sides of the Chinese central bank.  Scott Sumner has also added an excellent discussion of his own about what Chinese monetary policy is all about.  I am trying to assimilate the various responses […]

  17. Gravatar of TheMoneyIllusion » Governments manipulate currencies by saving – Let You Know Everything TheMoneyIllusion » Governments manipulate currencies by saving - Let You Know Everything
    6. November 2012 at 03:52

    […] Read this article: TheMoneyIllusion » Governments manipulate currencies by saving […]

  18. Gravatar of Scott Sumner Scott Sumner
    6. November 2012 at 08:18

    Costard, I never said bubbles couldn’t happen with stable NGDP growth, I said monetary policy would not be causing bubbles if you had stable NGDP growth. Some other factor would have caused it–perhaps lax regulation/moral hazard.

    The idea that China could have a signficant effect on the US natural rate of unemployment seems pretty far-fetched. Even Krugman doesn’t make that argument.

  19. Gravatar of Governments manipulate currencies by saving « Economics Info Governments manipulate currencies by saving « Economics Info
    7. November 2012 at 09:01

    […] Source […]

  20. Gravatar of rubin pham rubin pham
    7. November 2012 at 14:19

    chinese bashing is at an all time high here in the us.
    hopefully more people will read this article here.

  21. Gravatar of More on Currency Manipulation « Uneasy Money More on Currency Manipulation « Uneasy Money
    8. November 2012 at 20:48

    […] try to at least make a start on analyzing those effects in an upcoming post addressing the comments of Scott Sumner on my previous post. Share this:TwitterFacebookDiggLinkedInRedditStumbleUponGoogle […]

  22. Gravatar of Currency Manipulation: Is It Just About Saving? « Uneasy Money Currency Manipulation: Is It Just About Saving? « Uneasy Money
    12. November 2012 at 21:23

    […] my first discussion of currency manipulation, Scott Sumner responded with some very insightful comments of his own in which he pointed out that the current […]

  23. Gravatar of Bitcoin Faucet Rotator Blog Data visualization: The People's Bank of China balance sheet Bitcoin Faucet Rotator Blog Data visualization: The People's Bank of China balance sheet
    10. April 2015 at 06:55

    […] Glasner and Scott Sumner have posts on Chinese monetary policy. They both inquire about the People’s Bank of China […]

Leave a Reply