Saturday, September 3, 2011

Balance Sheet Recessions

On the Coordination Problem blog, Gerald O'Driscoll asked in the comment thread:

I want to address only one issue that you've raised. Once again, if Rogoff and Reinhart are correct, what you ask is impossible. Balance-sheet, financial crises of the type we have gone through lower the trend rate of growth for as long as 10 years.

Printing money cannot compensate for real changes. If you target NGDP, you will get more inflation and less growth.

If you want to be taken seriously, you must address their findings.

I did read This Time is Different.

My view is that the reduced real growth in the episodes they describe weren't due to "balance sheet recessions," but rather to rapid net capital outflows.

Generally, "balance sheet recession" supposedly occurs because net debt repayment reduces real expenditures on output. As households pay down debt, they reduce consumption expenditures. As firms pay down debt, they reduce investment expenditures.

This is just a version of the paradox of thrift. As long as monetary institutions prevent the development of an excess demand for money, (by perhaps printing enough,) households and firms can all pay down their debts while nominal expenditure is maintained.

It is possible that there could be some kind of adverse productivity shock associated with the repayment of debt, but there isn't necessarily such a shock. For example, perhaps some adverse productivity shock triggers a desire to pay down debt.

Or, perhaps, the process by which the market adjust to an increase in the supply of saving and decrease in the demand for investment--a lower interest rate--could require a change in the allocation of resources and substantial real adjustment costs.

If there is an adverse productivity shock, a stable growth path of NGDP doesn't prevent lower real growth. An adverse productivity shock results in the price level moving to a higher growth path (and so higher inflation for time,) and real output falling to a lower growth path, (and so, slower, or even negative, real growth for a time.)

If the demand for money relative to real income stays the same (constant velocity,) then a stable growth path of NGDP has the same consequences as a stable growth path for the quantity of money. In comparison to a target for the price level, an NGDP regime has higher inflation and a smaller decrease in real output when there is an adverse productivity shock.

So, what about Rogoff and Reinhart?

While I think that stabilizing the growth path of NGDP remains the best policy in response to a net capital outflow, it won't shield an economy from adverse effects for real growth.

Of course, the U.S. hasn't had a net capital outflow and so that isn't relevant as of yet. (I think the R&R research project was looking forward to just such an event and the actual crises we had was a square peg hammered into a round hole to help with marketing.)

If a country did have a net capital outflow (all of those foreign investors want to quit lending short and repatriate funds,) it is likely that some of the projects that were being funded will need to be abandoned, and so there will be a negative productivity shock. With an NGDP target, that would result in a higher domestic price level and a reduced volume of output.

Further, the exchange rate would fall and imports get more expensive. Using a CPI type deflator for measurement, domestic output is worth less. That could last for years.

Meanwhile, there needs to be a shift of labor and other resources from the investment projects that were funded with foreign funds to export and import competing industries. That suggests structural unemployment and other adjustment costs. It is like an adverse productivity shock.

And finally, the domestic natural interest rate is higher. That part of saving that was coming from the foreigners is no longer there. And while equity investors should take a big loss initially,there should be a shift away the share of income going to labor and towards the share going to capital. That suggests that nominal wages need to go to a lower growth path and nominal capital incomes to a higher growth path.

So, I think a net capital outflow is going to have an adverse impact on real incomes (and real wages) no matter what. While Rogoff and Reinhart didn't point to any episodes where NGDP was kept on a stable growth path despite a net capital outflow, there were lots of episodes where NGDP grew very rapidly and real output was still depressed.

While it is likely that some of this was due to going too far, with massive inflation being disruptive in and of itself, I have given all sorts of reasons why a net capital outflow will have adverse effects on real growth for a long time, and so, why NGDP targeting would result in a rising, or at least a higher, price level for some time.

But a net capital outflow isn't a balance sheet recession.

I have explained this may times. I believe that there was a speculative bubble in the housing market and resources have been frittered away building too many houses and too many capital goods specific to home production. Many workers have skills specific to housing construction.

Shifting away from housing production to other goods is a negative productive shock. Some of the productive capacity is irrelevant--specific to producing things that aren't as valuable as their opportunity costs.

If nominal expenditure had been kept on a stable growth path, the result would have a shift to a higher growth path of prices, and higher inflation for a time. Real output would have grown more slowly and perhaps even been reduced for a time. The unemployment rate would rise due to a higher natural unemployment rate--more structural unemployment.

However, the Federal Reserve allowed nominal expenditures to fall sharply to a point where they are now on a growth path approximately 14% below trend. It could well be that the Fed's efforts to make periodic changes in short term interest rates to stabilize expected inflation was not able to stabilize the growth path of nominal expenditures in the face of large scale deleveraging. The problem, however, was with the Fed's approach to policy.

Anyway, I believe that the decrease in real expenditure on output has resulted in a level of real output well below the depressed productive capacity of the economy. With NGDP targeting, that wouldn't have happened even with deleveraging. Instead, real output would be lower (perhaps 5% below trend rather than 10% below trend,) and the price level would be higher.

If a net capital outflow should develop, then the best option will still be a stable growth path of NGDP. Stabilizing the price level, the exchange rate, or the price of gold would make a bad situation worse.

Of course, it is possible that the net capital inflow that the U.S. enjoys could gradually shift to a net capital outflow, and real real wages and other real incomes could continue to grow without serious disruption.

However, I do agree with Rogoff and Reinhart. A rapid shift to a net capital outflow could happen here. And whatever unfortunate consequences it might have, the problem would not be a "balance sheet recession.


  1. Great post Bill.

    I completely agree. I personally have a very hard to understand the economic model, which is the foundation for the Balance Sheet Recession (BSR) argument.

    It seems like BSR guys either assume that a bubble lead to some misallocation of capital, which reduce long-term productivity growth. However, this is not addressed in “This time is Different” and in my view there is no real empirical documentation anywhere that bubbles reduces productivity growth. In fact take a country like the US. It is obvious that there is public finance problem, which needs to be addressed – and that will likely happen through a combination of spending cut and taxes increases as it is currently the case in a number of euro zone countries. Higher taxes obviously reduce productivity growth, but if the fiscal consolidation happens through spending cuts it is very likely to increase productivity growth. Similarly is the private sector increases savings it should increase productivity growth.

    Another BSR model could be an ultra Keynesian model where wage and prices are very rigid so the increase in private and government savings leads to a drop in aggregate demand and given the price and wage rigidities lead to years of low growth. However, in such a model monetary policy would actually solve the problem.

    I tend to agree that we need to look at capital outflows and more importantly on the funding of current account deficits. In most Emerging Markets the problems is an excess reliance on foreign currency funding. South Korea had debt in US dollars in 1997 and could not print dollars. Another example is a number of Central and Eastern European countries. Take Hungary. Here something like 60% of all household debt is foreign currency denominated (mostly in Swiss francs). As a result the Hungarian central bank is naturally very reluctant to ease monetary policy, as a weaker forint would create serious troubles for households funded in foreign currency, which undoubtedly could threaten financial stability.

    Hence, I think that foreign currency funding is the key story in most Balance Sheet Recessions. Furthermore, the empirical evidence that growth is weak for a long period following bubbles bursting also have to be seen in the light of central banks becoming hyper sensitive to bubble risks. A good example is Japan. Because Japan had a bubble Bank of Japan has become overly hawkish because it is afraid of repeating former mistakes. The Federal Reserve and ECB undoubtedly suffers from the same bubble fears. To me it is somewhat paradoxical most central banks today are nearly obsessed with the risk of bubbles, while very few central banks took the bubble risk serious back in 2005-7.

  2. As an counter example to the Balance Sheet Recession story one should mention the Czech Republic. Looking at Czech NGDP it looks as if there is balance sheet recession in the Czech economy. However, the Czech Republic - unlike the majority of Central and Eastern European economies - does not have a balance sheet problem: There is basically no foreign currency funding of households and companies, public and private debt is low and there never was any signs of a property market bubble in the Czech economy. So there can't be a BSR explanation of the very low growth in Czech NGDP. The explanation obviously is overly tight monetary policy (the central bank thinks monetary policy is loose because the key policy rate is 0.75%). Velocity has collapsed despite there is no balance sheet problems - so it is obvious an example of excess demand for money...

  3. You nicely undermine the notion of a "balance sheet recession," and helpfully offer net capital outflows as the causes of the phenomena Rogoff and Reinhart point to. O'Driscoll wrote: "If you want to be taken seriously, you must address their [R&R's] findings." I think you have filled the bill!
    You also wrote: "I think the R&R [Rogoff and Reinhart] research project was looking forward to just such an event [a net capital outflow] and the actual cris[i]s we had was a square peg hammered into a round hole to help with marketing." Touché!
    But it is jarring to read: "I believe that there was a speculative bubble in the housing market . . . ." A "speculative bubble" is just an investment that turns out badly. In 2007, and again in 2010, there was a "speculative bubble" in BankAmerica stock; in 1998 there was a "speculative bubble" in Ryan Leaf's talents for football. So *obviously* there was a "speculative bubble" in housing in 2006: no need for "I believe."

  4. I think these issues rises time and time again, it’s important to get proper understanding of these things and only then could be making good profits, but without that we will struggle. I am fairly lucky to be with OctaFX, it’s only thanks to them that I am able to work out nicely in any situation and that’s through their epic daily market news and analysis service which is easy to follow yet highly effective, so really cool that is.