By Michael Pettis. Cross posted from Michael Pettis’ China Financial Markets
One of the reasons that it is been so hard for a lot of analysts, even trained economists, to understand the imbalances that were at the root of the current crisis is that we too easily confuse national savings with household savings. By coincidence there was recently a very interesting debate on the subject involving several economists, and it is pretty clear from the debate that even accounting identities can lead to confusion.
The difference between household and national savings matters because of the impact of national savings on a country’s current account, as I discuss in a recent piece in Foreign Policy. In it I argue that we often and mistakenly think of nations as if they were simply very large households. Because we know that the more a household saves out of current income, the better prepared it is for the future and the more likely to get rich, we assume the same must be true for a country. Or as Mr. Micawber famously insisted:
Annual income twenty pounds, annual expenditure nineteen six, result happiness. Annual income twenty pounds, annual expenditure twenty pound ought and six, result misery.
But countries are not households. What a country needs to get wealthier is not more savings but rather more productive investment. Domestic savings matter, of course, but only because they are one of the ways, and probably the safest, to fund domestic investment (although perhaps because they are the safest, investment funded by domestic savings can also be misallocated for much longer periods of time than investment funded by external financing).
Saving in itself, however, does not create wealth. It is productive investment that creates wealth. Domestic savings simply represent a postponement of consumption.
In a closed economy, total savings is equal to total investment or, to put it differently, whatever we don’t consume we invest (if we produce something that we neither consume nor invest, we effectively write its value down to zero, so the balance remains). In an open economy, if a country saves more than it invests it must export the excess savings. It must also export the excess production.
Notice that by definition if a country saves more than it invests, total consumption plus total savings must be greater than total consumption plus total investment. The former is the sum of the goods and services it creates, whereas the latter is the sum of goods and services it absorbs. That country, in other words, supplies more goods and services than it absorbs, and so it must export the excess.
What is more, by exporting excess savings, the country is providing the funding to foreigners to purchase its excess production. This is why the current account and the capital account for any country must always add up to zero.
In the late 19th Century, as I discuss in my most recent book, economists like John Hobson in the UK and Charles Arthur Conant in the US noticed that the rich countries of the west were exporting large amounts of savings abroad – mostly to what were later called by the dependencia theorists of the 1960s the peripheral nations. Hobson and Conant argued that the reason for this excess savings had to do with income inequality. As more and more wealth is concentrated into the hands of fewer people, consumption rises more slowly than production, largely because the wealthier a person gets, the smaller the share he consumes out of his income. Notice that because savings is simply total production of goods and services minus total consumption, this forces up the national savings rate.
This was a very important insight. Excess savings, they pointed out, was not a result of old-fashioned thrift but rather a consequence of structural distortions in the economy. The consequence of this “thrift”, furthermore, was not greater wealth but rather structural imbalances in the global economy.
In a closed economy there are four ways of resolving the imbalances caused by an increase in the savings rate. First, and most obviously, investment can rise by the same amount.
The private sector, however, may be reluctant to increase investment if it believes the consumption share is declining over the long term, in which case the government can sponsor the increase in investment, for example in infrastructure, so that savings and investment balance at a higher rate. This is sustainable only as long as there are productive investments that can be made, but as consumption declines, the reason for investing should decline too. The purpose of investment today after all is to create consumption tomorrow.
The second way to resolve the imbalance is if the government or the labor unions take steps to redistribute income downwards. As middle class and poor households retain a greater share of total GDP, consumption will automatically rise relative to production (the savings rate declines), even if middle class and poor households save a greater share of their higher income. The savings rate declines to the point at which savings and investment are once again balanced domestically and everything a country makes it consumes or invests.
The third way to resolve this in a closed economy – albeit only temporarily – is to fund a consumption boom among the not-so-rich. How would this work? One way might be, as Conant discusses extensively, that as savings grow faster than opportunities for productive investment in infrastructure and production capacity, more and more of the savings of the wealthy go into speculative investments that drive up asset prices – homes, stocks and bonds. As asset prices rise, households feel richer and they begin to take advantage of the abundance of savings to borrow for consumption, and borrowing is just negative savings.
As home prices or the value of investment portfolios rise there is likely to be not just an increase in consumption but also an increase in investment in new housing. As both of these happen, the reduction in consumption caused by rising income inequality is matched by the increase in consumption caused by credit-fueled purchasing and an increase in housing investment, and once again savings and investment can balance domestically. We saw this happen in the US and in the peripheral countries of Europe in the run-up to the 2007-09 crisis.
The fourth way to resolve the savings imbalance in a closed economy is to force up unemployment (this is what Karl Marx said would eventually happen). As income inequality grows, and so consumption grows more slowly than production, companies are forced to cut production and fire workers. Fired workers of course produce nothing, but they still consume, either out of savings, welfare payments, or handouts from friends and families. This causes total savings to drop so that once again it balances investment, but of course in an economy with rising unemployment, profits are likely to drop, and with lower profits comes reduced investment, so more workers need to be fired and the process can become self-reinforcing.
Open economies have another option
In a closed economy there really aren’t many other ways to balance savings and investment if structural factors force up the savings rate. But we do not live in closed economies. Most of us live in open economies (although the world itself is a closed economy), so there is actually a fifth way to resolve domestic savings imbalances, and this is what Hobson and Conant described as the root source of late 19th Century imperialism.
If domestic savings rates are so high that the country cannot invest it all profitably, it can export those savings, which means automatically that it imports foreign demand for its excess production. Its net export of savings (less net returns on earlier investment) is exactly equal to its net export of goods and services.
In an open economy, in other words, a country’s total savings matters because to the extent that it exceeds investment, it must be exported, and it must result in a current account surplus. Here is where the confusion so many analysts, including economists, have about the difference between national and household savings. Household savings represent the amount out of household income that a household chooses not to consume, and so can be affected by cultural or demographic factors, the existence and credibility of a social safety net, the sophistication of consumer finance, and so on.
The national savings rate, on the other hand, includes not just household savings but also the savings of governments and businesses. It is defined simply as a country’s GDP less its total consumption. While the household savings rate may be determined primarily by the cultural and demographic preferences of ordinary households, the national savings rate is not. Indeed in some cases the household share of all the goods and services a country produces, which is primarily a function of policies and economic institutions, is the main factor affecting the national savings rate.
National savings, in other words, may have very little to do with household preferences and a lot to do with policy distortions. In China, which has by far the highest savings rate in the world, part of the reason for the high national savings rate of course is that Chinese households save a relatively high proportion of their income.
But while China’s savings rate is extraordinarily high, the Chinese household savings rate is merely in line with those of similar countries in the region, and in fact lower than some. Chinese households are not nearly as thrifty as their national savings rate implies. Why, then, is China’s savings rate so extraordinarily high?
The main reason, as I have discussed many times and which now has pretty much become accepted as the consensus among China specialists, is not so much income inequality (although this is certainly a problem in China) but rather the very low household income share of GDP. At roughly 50% of GDP, Chinese households retain a lower share of all the goods and services the country produces than households in any other country in the world.
This is a consequence of policies Beijing put into place many years ago that goose GDP growth by constraining the growth in household income. As a result of these policies, the household share of China’s total production of goods and services has been falling for thirty years, and fell especially sharply in the past decade. It isn’t surprising, consequently, that as households earn a declining share of what China produces, they also consume a declining share. Because savings is simply GDP less total consumption, and most consumption is household consumption, the fall in the household income share of GDP is the obverse of the rise in China’s extraordinarily high savings rate.
Many factors explain this very low household income share in China, including most importantly financial repression, whose characteristics typically include artificially low deposit rates, which, by reducing the amount of money that a saver should earn on his bank deposit, transfers part of his income to borrowers, who are able to borrow very cheaply. In China, this implicit transfer is extremely high, perhaps 5 percent of China’s GDP or more.
Of course the more money that is transferred in this way, the less disposable income the household depositor has, and so he is forced to reduce both his nominal savings and his nominal consumption. We cannot easily predict how this reduced interest rate will affect the household savings rate, but it is pretty easy to figure out how it will affect the national savings rate. If the transfer is substantial, it will reduce the share of GDP retained by households. Unless households reduce their savings rate by more than the reduction in the household share of GDP, it must automatically force up the national savings rate.
Confusing thrift with inequality
China’s extraordinarily high national savings rate, in short, is a function primarily of the extraordinarily low household share of GDP. Even economists who really should know better manage to make some fairly impressive mistakes when they discuss Chinese and other savings imbalances, mostly because their understanding of savings can be hopelessly confused. For a typical example, consider a piece Raman Ahmed and Helen Mees, published last year called, “Why do Chinese households save so much?”
In the article the authors try to address the causes of China’s high savings rate, but they do so by thoroughly confusing national savings with household savings. For example they set out trying to prove that financial repression has no impact on China’s savings rate, but because they fail to understand that financial repression does this by reducing the household share of income, and not necessarily by reducing the household savings rate, they find:
China’s monumental savings rate is a popular topic of for policy discussion. It has been blamed for the global financial crisis, currency wars, and the ensuing Great Recession. But what explains the high savings rate?
…Although the savings rate varies significantly per income group, with the lowest income group’s savings rate in urban areas in the single digits and the highest income group’s savings rate at almost 40% of disposable income, we do not find evidence that income inequality as such is a motive for households to save a larger portion of their income, as Jin et al. (2010) have suggested. It would have been the Chinese version of ‘keeping up with the Joneses’, albeit that ‘keeping up with the Wangs’ would not have involved conspicuous consumption but rather conspicuous saving. This would have instigated higher savings rates across the board of deciles of savings rates, with the strongest effect on low-income households. However, using urban data on household savings rates and income inequality from 1985-2009, we do not find this effect present.
There is no evidence that the household savings rate in China is high because of low deposit rates, as Michael Pettis (2012) has asserted time and again, which would indicate that the income effect of lower deposit rates trumps the substitution effect of lower deposit rates. The coefficient of the deposit rate has alternating signs, but is insignificant in every single estimate.
The article purports to discuss what they refer to as China’s “monumental savings rate”, which, according to the authors, has been blamed for the global financial crisis, currency wars, and the ensuing Great Recession, but it focuses on the wrong savings rate. Chinese household savings are not by any definition “monumental”, and they most certainly did not cause the global financial crisis, nor does anyone seriously claim that they did. Chinese household savings rates are high, but not exceptionally high, and because household income is such a low share of GDP, Chinese household savings as a share of GDP, which is what really matters, are even lower than the household savings rate would imply. Chinese household savings are not the problem.
It is China’s national savings rate which is “monumental” and which drives China’s current and capital account imbalances, and the national savings rate is monumentally high because the national consumption rate (which consists mostly of the household consumption rate) is extraordinarily low. The authors have either confused national and household savings rates or they have failed to see that what matters is not the household savings rate but rather total savings. Aside from the fact that there is indeed evidence that Chinese savings are negatively correlated with interest rates, for example a 2011 study done by the IMF, the relationship is one of pure logic.
The important lesson from this article, aside from suggesting just how confused many economists are when it comes to understanding the source of global imbalances, is that national savings represent a lot more than the thriftiness of local households, and as such it has a lot less to do with household or cultural preferences than we think. In fact many factors affect the savings rate of a country, including demographics, the extent of wealth inequality, and the sophistication of consumer credit networks, but when a country has an abnormally high savings rate it is usually because of policies or institutions that restrain the household share of GDP.
This has happened not just in China but also in Germany. In the 1990s Germany could be described as saving too little. It often ran current account deficits during the decade, which means that the country imported capital to fund domestic investment. A country’s current account deficit is simply the difference between how much it invests and how much it saves, and Germans in the 1990s did not always save enough to fund local investment.
But this changed in the first years of the last decade. An agreement among labor unions, businesses and the government to restrain wage growth in Germany (which dropped from 3.2 percent in the decade before 2000 to 1.1 percent in the decade after) caused the household income share of GDP to drop and, with it, the household consumption share. Because the relative decline in German household consumption powered a relative decline in overall German consumption, German saving rates automatically rose.
Notice that German savings rate did not rise because German households decided that they should prepare for a difficult future in the eurozone by saving more. German household preferences had almost nothing to do with it. The German savings rate rose because policies aimed at restraining wage growth and generating employment at home reduced household consumption as a share of GDP.
As national saving soared, the German economy shifted from not having enough savings to cover domestic investment needs to having, after 2001, such high savings that not only could it finance all of its domestic investment needs but it had to invest abroad by exporting large and growing amounts of savings. As it did so its current account surplus soared, to 7.5 percent of GDP in 2007. Martin Wolf, in an excellent Financial Times article on Wednesday on the subject, points out that
between 2000 and 2007, Germany’s current account balance moved from a deficit of 1.7 per cent of gross domestic product to a surplus of 7.5 per cent. Meanwhile, offsetting deficits emerged elsewhere in the eurozone. By 2007, the current account deficit was 15 per cent of GDP in Greece, 10 per cent in Portugal and Spain, and 5 per cent in Ireland.
Employment policies and the savings rate
It is tempting to interpret Germany’s actions as the kind of far-sighted and prudent actions that every country should have followed in order to keep growth rates high and workers employed, but it turns out that these policies did not solve unemployment pressures in Europe, and this is implied in the second sentence of Martin Wolf’s piece. Germany merely shifted unemployment from Germany to elsewhere. How? Because Germany’s export of surplus savings was simply the flip side of policies that forced the country into running a current account surplus.
To explain, let us pretend that Europe consists of only two countries, Spain and Germany. As we have already shown, forcing down the growth rate of German wages relative to GDP caused the household income share of GDP to drop. Unless this was matched by a decision among German households to become much less thrifty, or a decision by Berlin to increase government consumption sharply, the inevitable consequence had to be a reduction in the overall consumption share of GDP, which is just another way of saying that the German national savings rate had to rise. During this period, by the way, and perhaps as a consequence of restraining wages, Germany’s Gini coefficient seems to have risen quite markedly, and the resulting increase in income inequality also affected savings adversely.
As German savings rose, eventually exceeding German investment by a wide margin, Germany had to export the difference, which its banks did largely by making loans into the rest of Europe, and especially those countries that were financially “shallower”. Declining consumption left Germany producing more goods and services than it could absorb domestically, and it exported excess production as the automatic corollary to its export of savings.
Of course the rest of the world had to absorb excess German savings and run the current account deficits that corresponded to Germany’s surpluses. This was always likely to be those eurozone countries that joined the monetary union with a history of higher inflation and currency depreciation than Germany – countries which we are here calling “Spain”. As monetary policy across Europe was made to fit German needs, which was looser than that required by Spain, and as German savings were intermediated by German banks into Spain, the result was likely to be higher wage growth, higher inflation, and soaring asset prices in Spain.
In fact this is exactly what happened. Spain and the other peripheral European countries all saw their trade deficits expand dramatically or their surpluses (many were running large surpluses in the 1990s) turn into large deficits shortly after the creation of the single currency as their savings rates shifted to accommodate German exports of its excess savings.
The way in which the German exports of savings were absorbed by Spain is at the heart of the subsequent crisis. As long as Spain could not use interest rates, trade intervention, or currency depreciation to block German exports, it had no choice but to balance the excess of German savings over investment. This meant that either its investment would have to rise or its savings would have to fall (or both).
Both occurred. Spain increased investment in infrastructure and in real estate (and less so in manufacturing, probably because German growth occurred at the expense of the manufacturing sectors in the rest of Europe), but it seems to have done both to excess, perhaps because of the sheer amount of capital inflows. After nearly a decade of inflows larger than any it had ever absorbed before, Spain, like nearly every country in history under similar circumstances, ended up with massive amounts of misallocated investment.
But this was not all. If the savings that Germany exported into Spain could not be fully absorbed by the increase in Spanish investment, the only other way to balance was with a sharp fall in Spanish savings. There are two ways Spanish savings could have fallen. First, as the Spanish tradable goods sector lost out to German competition, Spanish unemployment could rise and so force down the Spanish savings rate (unemployed workers still must consume).
Second, Spain could have reduced household savings voluntarily by increasing consumption relative to income. Higher Spanish consumption would cause enough employment growth in the services and real estate sectors to make up for declining employment in the tradable goods sector.
Not surprisingly, given the enormous optimism that accompanied the creation of the euro, the latter happened. As German money poured into Spain, helping ignite a stock and real estate boom, ordinary Spaniards began to feel wealthier than they ever had before, especially those who owned their own homes. Thanks to this apparent increase in wealth, they reduced the amount they saved out of current income, as households around the world always do when they feel wealthier. Together the reduction in Spanish savings and the increase in Spanish investment (in infrastructure and real estate) was enough to absorb the full extent of Germany’s export of excess savings.
But at what cost? The imbalance created within Europe by German policies to constrain consumption forced Spain into increasing consumption and boosting investment, much of the latter in wasted real estate projects (as happened in every one of the deficit countries that faced massive capital inflows). There are of course no shortage of moralizers who insist that greed was the driving factor and that Spain wasn’t forced into a consumption boom. “No one put a gun to their heads and forced them to buy flat-screen TVs”, they will say,
But this completely misses the point. Because Germany had to export its excess savings, Spain had no choice except to increase investment or to allow its savings to collapse, with the latter either in the form of a consumption boom or a surge in unemployment. No other option was possible.
To insist that the Spanish crisis is the consequence of venality, stupidity, greed, moral obtuseness and/or political short-sightedness, which has become the preferred explanation of moralizers across Europe begs the question as to why these unflattering qualities only manifested themselves after Spain joined the euro. Were the Spanish people notably more virtuous in the 20th century than in the 21st? It also begs the question as to why vice suddenly trumped virtue in every one of the countries that entered the euro with a history of relatively higher inflation, while those eastern European countries with a history of relatively higher inflation that did not join the euro managed to remain virtuous.
The European crisis, in other words, had almost nothing to do with thrifty Germans and spendthrift Spaniards. It had to do with policies aimed at boosting German employment, the secondary impact of which was to force up German national savings rates excessively. These excess savings had to be absorbed within Europe, and the subsequent imbalances were so large (because German’s savings imbalance was so large) that they led almost inevitably to the circumstances in which we are today.
For this reason the European crisis cannot be resolved except by forcing down the German savings rate. And not only must German savings rates drop, they must drop substantially, enough to give Germany a large current account deficit. This is the only way the rest of Europe can unwind the imbalances forced upon the region in a way that is least damaging to Europe as a whole. Only in this way can countries like Spain stay within the euro while bringing down unemployment.
But lower German savings don’t mean that German families should become less thrifty, only that the average German household should be allowed to retain a much larger share of what Germany produces. If Berlin were to cut consumption taxes, or cut income taxes for the lower and middle classes, or force up wages, total German consumption would rise relative to GDP and so national savings would fall – without requiring any change in the prudent behavior of German households.
To ask Spanish households to be more “German” by saving more is not only impractical in an economy with 25 percent unemployment (it is hard for unemployed workers to increase their savings), it is counterproductive. Lower Spanish consumption can only cause even higher Spanish unemployment, until eventually Spain will be forced to abandon the euro and so regain control of its ability to absorb or reject German imbalances. This abandonment of the euro will be driven by the political process, as those in the leadership (of both main parties) who refuse to countenance talk of leaving the euro lose voters to more radical parties until they, too, come around:
The latest opinion polls show the PP has lost 10 percentage points of support since Rajoy’s election in November 2011. Support for the main opposition group, the Socialists remains unchanged, while backing has been growing for smaller, more radical, parties. More than 68 per cent of Spaniards say the government is doing a bad or very bad job, while the latest official forecast shows that a quarter of the workforce will still be out of work three years from now.
An article in this week’s Economist suggests that Spain is indeed becoming more “German” and so that this is reason for hope:
Is Spain the next Germany? It may not feel like that to the 26% of Spaniards who are unemployed. GDP shrank by 0.8% in the fourth quarter of 2012. Yet in some ways, Spain resembles the Germany of a decade ago, when Gerhard Schröder brought in reforms to turn the sick man of Europe into its strongest economy. The efforts by Mariano Rajoy’s government to loosen labour laws and cut public spending are aimed at a German-style miracle.
…Joachim Fels, chief economist at Morgan Stanley, is one of several backers of the Germany theory. “Spain is doing a lot of the things Germany did ten years ago, but in a much shorter time span and tougher global conditions,” he says, pointing to falling labour costs, rising exports and booming Spanish car factories. But, he adds, “Spain becoming Germany is really a two- to four-year story.”
The global constraints
The problem with this argument may be, however, that the global conditions that allowed Germany to grow by exporting savings to Spain cannot be replicated. If Spain were to make its workers more competitive by reducing wage growth relative to GDP growth, it would implicitly be forcing up its savings rate to generate employment. To whom would Spain export those savings? The world is awash in excess savings, and unlike in pre-crisis days, there are no countries with booming stock and real estate markets willing to fund another consumption binge. This means that Spain and Europe are expecting to recover by exporting unemployment, but to whom?
In fact this is the great worry that Martin Wolf expresses n the conclusion to his article:
A big adverse shock risks turning low inflation into deflation. That would aggravate the pressure on countries in crisis. Even if deflation is avoided, the hope that they will grow their way out of their difficulties, via eurozone demand and internal rebalancing, is a fantasy, in the current macroeconomic context.
That leaves external adjustment. According to the IMF, France will be the only large eurozone member country to run a current account deficit this year. It forecasts that, by 2018, every current eurozone member, except Finland, will be a net capital exporter. The eurozone as a whole is forecast to run a current account surplus of 2.5 per cent of GDP. Such reliance on balancing via external demand is what one would expect of a Germanic eurozone.
If one wants to understand how far the folly goes, one must study the European Commission’s work on macroeconomic imbalances. Its features are revealing. Thus, it takes a current account deficit of 4 per cent of GDP as a sign of imbalance. Yet, for surpluses, the criterion is 6 per cent. Is it an accident that this happens to be Germany’s? Above all, no account is taken of a country’s size in assessing its contribution to imbalances. In this way, Germany’s role is brushed out. Yet its surplus savings create huge difficulties when interest rates are close to zero. Its omission makes this analysis of “imbalances” close to indefensible.
The implications of the attempt to force the eurozone to mimic the path to adjustment taken by Germany in the 2000s are profound. For the eurozone it makes prolonged stagnation, particularly in the crisis-hit countries, highly likely. Moreover, if it starts to work, the euro is likely to move upwards, so increasing risks of deflation. Not least, the shift of the eurozone into surplus is a contractionary shock for the world economy. Who will be both able and willing to offset it?
The eurozone is not a small and open economy, but the second-largest in the world. It is too big and the external competitiveness of its weaker countries too frail to make big shifts in the external accounts a workable post-crisis strategy for economic adjustment and growth. The eurozone cannot hope to build a solid recovery on this, as Germany did in the buoyant 2000s. Once this is understood, the internal political pressures for a change in approach will surely become overwhelming.
As long as it is part of the euro Spain has no choice but to respond to changes in German savings rates. There is nothing mysterious about this process. It is simply the way the balance of payments works, and thrift has nothing to do with it. If Germany does not take steps to force down its savings rate by increasing the household share of GDP, then either all of Europe becomes like Germany, in which case growth slows to a crawl and some other country – maybe the US? – will be forced to resolve Europe’s demand deficiency either through higher unemployment or through higher debt, or Europe must break apart to free Spain and the other peripheral countries from German savings imbalances.
I don’t imagine the rest of the world can absorb demand deficiency from a Germanic Europe, and if Europe tries to force it the result will almost certainly be an eventual collapse in trade relations, so either Germany rebalances or Europe breaks apart. It is hard for me to see many other options.