Why is Piketty so certain about things which are probably completely wrong?
10 July 2014
SUGGESTED
Housing and Planning
Healthcare
Housing and Planning
Scott Sumner produced an excellent blog post on Piketty recently. It begins by quoting Piketty:
‘In my view, there is absolutely no doubt that the increase of inequality in United States contributed to the nation’s financial instability. The reason is simple: one consequence of increasing inequality was virtual stagnation of the purchasing power of the lower and middle classes in the United States, which inevitably made it more likely that modest households would take on debt.’
Sumner’s blog really should be read in full as it reveals other factual inaccuracies in Piketty’s book that are then used as the basis for his economic judgements. Those factual inaccuracies all seem to suggest a particular set of political prejudices that then colour normative judgements.
For reasons Sumner identifies, the quotation above is an extraordinary statement from an economist. It should be noted that Piketty says ‘there is absolutely no doubt’ that the trends he described contributed to the crash. Yet we would, as Sumner points out, expect the opposite trend in debt and savings if incomes stagnate. If real incomes stop rising after a period of strong growth, and are expected to be flat in the future, according to the permanent income hypothesis we would expect saving to rise. A previously lower level of saving predicated upon expectations of rising incomes would no longer be justified.
Of course, left-leaning economists keep telling us that people do not behave rationally all the time – though they should be careful not to assume that people necessarily behave irrationally all the time. And, interestingly, on this side of the Atlantic, things were rather different.
In Britain, inequality fell. Whilst in the US, there was a rise in the Gini coefficient from 0.35 to 0.38 between 1990 and 2010, in the UK there was a small fall in the Gini coefficient from 0.35 to 0.34. In the UK, not only were the incomes of the poor rising relative to those of the rich, there was good reason for individuals to expect high future economic growth in general (even if many economists were sceptical). Average annual growth was nearly 2.5 per cent from 2000-2008 and Gordon Brown famously raised the official estimate of the sustainable growth rate.
But, despite these very different circumstances, the outcome was the same. In both the UK and the US, the savings ratio fell by about two-thirds. Looking at the data over other periods makes no difference to the conclusion about income growth, savings or inequality. Indeed, there were discussions at Shadow Monetary Policy meetings around 2007 about whether the low savings rate was a concern or whether it was a natural reaction to expectations of higher income growth in accordance with permanent income hypothesis.
There is no shortage of potential alternative explanations for the rise in household borrowing in both the UK and the US. Four that immediately come to mind are:
1. The rise in house prices led households’ net wealth to increase so they believed they could borrow more or save less.
2. As mentioned by Sumner, the large rise in savings in Asia lowered world real interest rates and therefore lowered savings elsewhere.
3. As Austrian business cycle theory would suggest, loose monetary policy led to low costs of borrowing and reduced saving and/or increased borrowing.
4. There was moral hazard prevailing right through the financial system in the US which artificially reduced the risks to both borrowers and lenders.
It has to be said, that the similar experience of the US and UK in relation to debt and saving in the face of completely different experiences of income and inequality data do not provide a very firm basis for Piketty being ‘absolutely certain’ that inequality contributed to the crash. If Piketty were not a respected economist, one might assume that he held a prior left-leaning view around which he is trying (not very successfully) to fit to the evidence.
A further comment on this issue has come from Frances Coppola, and her points do need to be addressed. She suggested – in elaborating on Piketty’s own explanation and attacking Sumner – that the increase in inequality in the US led to a large pool of private savings held by the rich and by corporate entities that was then used to fund borrowing by those whose incomes had stagnated. Coppola argues that, in boom times, savers prefer not to buy real assets but to lend more to individuals because they become less risk averse and individuals pay higher interest rates. However, this does not address the fundamental point. If a rich individual is lending to a poorer individual, that poorer individual must want to borrow. But, why would poor individual wish to borrow if his income is not expected to increase? And why were the savings of the rich not invested abroad? Coppola argues that Piketty (and Coppola) are taking a sector balance approach to the problem rather than focusing on individual behaviour. But to take such an approach is not economics: sector balances are not handed down from heaven but depend on individual behaviour (and the behaviour of the government and the corporate sector). Coppola argues that the excess savings of the rich had to be invested at home because the US had a trade deficit and therefore could not be a net accumulator of overseas assets. However, the trade deficit is a function of the supply of and demand for goods, services and capital and, if the supply of US capital to overseas borrowers increases, the exchange rate will fall and the balance of trade adjust in the other direction. The US trade deficit is not some kind of fixed constant around which everything else must vary.
For all the problems of the neo-classical/new-Keynesian synthesis, old-fashioned hydraulic Keynesianism, which simply takes economic variables as given when the causes of changes in those variables are the very subject of economic science, is no replacement.
Update: Scott Sumner responds to Philip Booth’s post here.
‘In my view, there is absolutely no doubt that the increase of inequality in United States contributed to the nation’s financial instability. The reason is simple: one consequence of increasing inequality was virtual stagnation of the purchasing power of the lower and middle classes in the United States, which inevitably made it more likely that modest households would take on debt.’
Sumner’s blog really should be read in full as it reveals other factual inaccuracies in Piketty’s book that are then used as the basis for his economic judgements. Those factual inaccuracies all seem to suggest a particular set of political prejudices that then colour normative judgements.
For reasons Sumner identifies, the quotation above is an extraordinary statement from an economist. It should be noted that Piketty says ‘there is absolutely no doubt’ that the trends he described contributed to the crash. Yet we would, as Sumner points out, expect the opposite trend in debt and savings if incomes stagnate. If real incomes stop rising after a period of strong growth, and are expected to be flat in the future, according to the permanent income hypothesis we would expect saving to rise. A previously lower level of saving predicated upon expectations of rising incomes would no longer be justified.
Of course, left-leaning economists keep telling us that people do not behave rationally all the time – though they should be careful not to assume that people necessarily behave irrationally all the time. And, interestingly, on this side of the Atlantic, things were rather different.
In Britain, inequality fell. Whilst in the US, there was a rise in the Gini coefficient from 0.35 to 0.38 between 1990 and 2010, in the UK there was a small fall in the Gini coefficient from 0.35 to 0.34. In the UK, not only were the incomes of the poor rising relative to those of the rich, there was good reason for individuals to expect high future economic growth in general (even if many economists were sceptical). Average annual growth was nearly 2.5 per cent from 2000-2008 and Gordon Brown famously raised the official estimate of the sustainable growth rate.
But, despite these very different circumstances, the outcome was the same. In both the UK and the US, the savings ratio fell by about two-thirds. Looking at the data over other periods makes no difference to the conclusion about income growth, savings or inequality. Indeed, there were discussions at Shadow Monetary Policy meetings around 2007 about whether the low savings rate was a concern or whether it was a natural reaction to expectations of higher income growth in accordance with permanent income hypothesis.
There is no shortage of potential alternative explanations for the rise in household borrowing in both the UK and the US. Four that immediately come to mind are:
1. The rise in house prices led households’ net wealth to increase so they believed they could borrow more or save less.
2. As mentioned by Sumner, the large rise in savings in Asia lowered world real interest rates and therefore lowered savings elsewhere.
3. As Austrian business cycle theory would suggest, loose monetary policy led to low costs of borrowing and reduced saving and/or increased borrowing.
4. There was moral hazard prevailing right through the financial system in the US which artificially reduced the risks to both borrowers and lenders.
It has to be said, that the similar experience of the US and UK in relation to debt and saving in the face of completely different experiences of income and inequality data do not provide a very firm basis for Piketty being ‘absolutely certain’ that inequality contributed to the crash. If Piketty were not a respected economist, one might assume that he held a prior left-leaning view around which he is trying (not very successfully) to fit to the evidence.
A further comment on this issue has come from Frances Coppola, and her points do need to be addressed. She suggested – in elaborating on Piketty’s own explanation and attacking Sumner – that the increase in inequality in the US led to a large pool of private savings held by the rich and by corporate entities that was then used to fund borrowing by those whose incomes had stagnated. Coppola argues that, in boom times, savers prefer not to buy real assets but to lend more to individuals because they become less risk averse and individuals pay higher interest rates. However, this does not address the fundamental point. If a rich individual is lending to a poorer individual, that poorer individual must want to borrow. But, why would poor individual wish to borrow if his income is not expected to increase? And why were the savings of the rich not invested abroad? Coppola argues that Piketty (and Coppola) are taking a sector balance approach to the problem rather than focusing on individual behaviour. But to take such an approach is not economics: sector balances are not handed down from heaven but depend on individual behaviour (and the behaviour of the government and the corporate sector). Coppola argues that the excess savings of the rich had to be invested at home because the US had a trade deficit and therefore could not be a net accumulator of overseas assets. However, the trade deficit is a function of the supply of and demand for goods, services and capital and, if the supply of US capital to overseas borrowers increases, the exchange rate will fall and the balance of trade adjust in the other direction. The US trade deficit is not some kind of fixed constant around which everything else must vary.
For all the problems of the neo-classical/new-Keynesian synthesis, old-fashioned hydraulic Keynesianism, which simply takes economic variables as given when the causes of changes in those variables are the very subject of economic science, is no replacement.
Update: Scott Sumner responds to Philip Booth’s post here.
5 thoughts on “Why is Piketty so certain about things which are probably completely wrong?”
Comments are closed.
SHARE
“Why would a poor individual wish to borrow if his income is not expected to increase?” That is not actually what either Sumner or Piketty say.
Here is Piketty:
“one consequence of increasing inequality was virtual stagnation of the purchasing power of the lower and middle classes in the United States”.
And here is Sumner:
“Why would stable real wages lead to more debt? More specifically, why would it “inevitably” lead to more debt? If my wages were stagnant I certainly wouldn’t react by taking on more debt, rather I’d borrow more if I expected my income to rise. Is there a theory here?”
They are not talking about the same thing. Sumner is talking about individual expectation of future income rises. Piketty is talking about what actually happened to low and middle incomes in aggregate.
In my takedown of Sumner I discussed the expectations problem:
“Why would people borrow when they don’t expect their incomes to rise? This is indeed a good question. But it is not what Piketty actually says. He notes that the purchasing power of low-to-middle income people stagnated. But he says nothing about their expectations. People expect their incomes to rise over the course of their working life. This expectation is near-universal and has nothing to do with short-term economic conditions. Even when average incomes are falling, individuals still generally expect their incomes to rise over the medium term.”
(full post is here – it is not the same as the one you have linked to: http://coppolacomment.blogspot.co.uk/2014/07/sumner-on-piketty.html)
So it is entirely possible, and indeed likely, that people might borrow in anticipation of FUTURE income rises when their current incomes are actually stagnating.
Regarding the sectoral balances approach. The US trade deficit is pretty intractable, largely because the two major surplus countries – China and Germany – do not have currencies that float with respect to the USD. Germany uses the Euro, which does float, but the Euro is persistently undervalued relative to fundamentals in Germany because of the presence of weaker countries in the union. If the currency cannot adjust, then neither the trade deficit nor the capital surplus can correct unless unit labour costs fall, which means very significant falls in wages and employment costs. I’m very happy to discuss the microeconomic factors that support this situation, and indeed I am planning to do so in another post, following extensive twitter discussions about this with a variety of people. But that is the macroeconomic context. Your statement that it is “not economics” suggests you have a rather narrow view of what constitutes “economics”.
thank you for your comment, Frances. On the contrary, I have a very, very wide view of what constitutes economics, but I think one has to go back to human behaviour in the economic sphere rather than assume that something just happens to be fixed and given; and it is very clear that I was saying that the view that sector imbalances just “exist” is not economics. You have elaborated now and thus we are back in the sphere of economics! So let’s move on from what might be regarded as petty point scoring (started by me).
On the issue of borrowing and future expectations etc.,…those who run the inequality/incomes stagnation line (which is disputed I should add, but if one disputes it then one disputes the inequality figures too) then the argument goes that real incomes among the least well off have stagnated in the US since 1970. This is not short term and should be a long enough period of affect expectations. But, in any case, if those expectations of increased incomes were validly held then the borrowing would be rational (which you would accept). If they were validly held and subsequently validated then a lot of people would probably stop worrying about the poverty issue.
Regarding the sector balances, then effectively the argument seems to have moved on to one that is often used by central bankers and others that the financial crash was really caused by this problem and not by inequality. However, whilst it may be true that many of the factors you point to (e.g. currency values) are fixed or managed, others are not. China will not be able to supress inflation forever if its currency is under-valued. As you say, Germany is part of the euro and there will be internal inflation in Germany if its economy is out of equilibrium with the others in the euro zone. If there are positive balances in the US private, corporate and government sectors then the dollar can float down easily enough. However, the fact is that, taking the government, corporate and household sectors together the US economy has generally been a net dis-saver. Such a country will import capital and its exchange rate will rise. Roughly speaking, both the government and household sectors were net borrowers to the tune of 2.5 per cent of GDP. Unsurprisingly, the US trade deficit was 5 per cent of GDP. China and Germany both had the opposite situation. If the US private sector had been in surplus, everything else would not have remained the same. I struggle to understand how a robust theory can have the balance of payments deficit driven only by trade with net savings somehow being a residual and not affecting the exchange-rate setting (or in a fixed exchange rate model) monetary and inflation process. It just assumes away a whole lot of important economic variables (which is why I used the tendentious phrase “not economics”)
http://www.prosper.org.au/2009/01/28/ineffective-demand/
As is illustrated in the above link, Riccardo’s Law of Rent predicts that as GDP rises, the return to wages and interest fall, and the return to rent rises.
No need for any convoluted Neo-Classical economic explanations. Neo-Classical Economics was formulated precisely to obscure people from see the very simplest economic truths. Job done I’d say.
Borrowing in expectation of future income rises is clearly rational. The question is whether, given the (claimed) stagnation of incomes since 1970, that expectation is or was rational. I would argue that it is. Even if incomes in aggregate are stagnating, individual incomes for prime-age workers tend to rise as they gain skills and experience:this is balanced by individual incomes tending to fall over the age of 50 and by women dropping down to lower income levels when they have children. The tendency of economists to speak of “households”, ignoring their composition, masks this last effect in particular.The problem, of course, is that for many people their rational expectations (or if you prefer, hopes and dreams) of better jobs and higher incomes have been dashed by the reality of techological change, offshoring, and since the crisis recession, stagnation and unemployment. So actually I didn’t disagree with Scott’s observation that borrowing would be done in anticipation of higher income. I disagreed with his (and your) reading of Piketty.
Regarding the sectoral balances, I accept your point about low savings in the US. I would add that Americans in general (not just those on low incomes) were encouraged to borrow against property for consumption by President Bush after 9/11: had this not resulted in a consumption boom, the US government’s borrowing would have been far higher (I think I said this in the post), not least because of the Afghanistan and Iraq wars. We could say that the tax revenue generated from the consumption boom went to finance the war on terror.
You criticise me for treating net saving as a residual. I didn’t, actually, but I perhaps did not give it as much weight as you would like. But equally, you can’t posit a balance of trade theory driven entirely by net saving while ignoring the effects of outright financial repression in both China and Germany, either. If the US increased its net saving under these circumstances the trade deficit and capital surplus would indeed fall, but that wouldn’t necessarily result in an increase in US exports. It might simply mean a reduction in everyone’s trade – and trade DOES matter. It is global trade above all that pulls people out of poverty. I don’t wish to see it reduce.
thank you, Frances, very sensible thoughts