Escape From Balance Sheet Recession - Richard Koo

Amazon link

  • BSR = firms minimizing debt at the expense of maximizing profits

  • BSR characterized by credit extension being a fraction of general willingness to lend, i.e. a shortage of private borrowers

  • QE distorts bond yield signaling function, and fuels asset bubbles that does little for the real economy in the long-run – it is also extremely expensive

A month earlier I first encountered the term “balance sheet recession” (BSR) in A. Turner’s Between Debt and the Devil, a great, accessible introduction to the concept of debt and the pernicious problem it represents today. In the footnotes, Turner mentioned this book, the reading of which I found to be at times tedious and repetitive (it’s translated – not too well obviously – from the original Japanese) but makes some bold and most importantly, substantiated by data, claims.


that we’re in a BSR. What is a BSR? Koo defines it as a situation where economics agents seek to minimize debt, even at the expense of maximizing profits. Econs 101 uses the latter as an assumption to underpin economic theory and policy, including recovery policy… do you see the dilemma here? Since the Great Recession (the causes of which have been covered well and via a story narrative in Lewis’ The Big Short), governments around the world have turned to quantitative easing (QE – which I might cover later in the future, but for now will just touch briefly), the buying of government bonds in the open market, i.e. from commercial banks, thereby increasing bond prices, and ultimately driving down bond yields and increasing commercial bank reserves. The goal of QE is that with lower bond yields and higher bank reserves, banks will lend more. Essentially it’s monetary policy based on a profit maximization assumption. 

However, as banks seek to minimize debt in a BSR at the cost of profits, increases in the monetary base wouldn’t translate into credit extension—and this is exactly what has happened:

(366% increase in monetary base has resulted in a 5% increase in credit extension, and 46% increase in M2 )

(366% increase in monetary base has resulted in a 5% increase in credit extension, and 46% increase in M2 )

Why? Well with economic agents (firms and households) seeking to minimize debt, the lowest interest rates in history haven’t been enough to tempt them to borrow – they’ve been net savers for 6 years. With the recent financial crisis still in the back of their minds, consumers around the world simply aren’t reverting back to their old ways. So we’re in a BSR, characterized by an absence of private borrowers. 

At this point I’d like to highlight how I was also confused over the concept of a BSR. Like, how come we haven’t heard of this before; how can economists have been so wrong? Koo notes that the West, out of fear and conceit, won’t admit it’s going through what Japan did in the 90’s, disguising today’s BSR as a recession borne of structural reasons, and that the Chicago School/modern interpretation of the New Deal is that it worked thanks to increases in the monetary base (however those increases never translated into increased credit extension, similarly to today).  So with the two historical examples of BSR refuted as even being BSRs, it comes as no surprise that little is still known about BSRs and accordingly we’ve reacted wrongly.

  • two types of recession: structural and BSR

  • BSR = debt minimization = agents don’t react to near-zero interest rates, they simply won’t borrow

  • absence of private borrowers


monetary policy is ineffective at combatting a BSR, like throwing strawberry jelly at a dragon.  Wait what does an absence of private borrowers mean? By definition it means there are excess savings. And this is where the big problem is. Savings are a withdrawal from the income stream of an economy, and if they’re not re-injected via investment (i.e. borrowing, or FDI), the economy shrinks by a multiplied effect

These excess savings therefore represent the deflationary gap, the different between potential and actual GDP. And while monetary policy can make the bait/borrowing more attractive, if the fish aren’t hungry, then they simply won’t bite no matter how attractive the bait/rates are. That’s the strawberry jelly.

  • absence of private borrowers = excess savings => deflationary spiral

  • monetary policy is ineffective at fixing a BSR


fiscal policy is the knight in shining armor. Koo notes that Japan-bashers have criticized Japanese fiscal stimulus since the 90’s as ineffective, implying that they believe the economy could have achieved 0% growth (what it has achieved) without stimulus. The Japan 1991 crash was accompanied by a 87% collapse in the price of commercial real estate, the closest comparable modern economy recession is the Great Depression which wiped out 46% of US GDP and took 30y for rates to normalize and the largest war effort to get in the economy back into growth mode. So in that context, Japan’s fiscal stimulus worked, and extremely well at that.

The supposed chink in the knight’s armor is the economic concept of “crowding out”, the idea that when the government funds large public spending projects to create demand, it’s siphoning funds from what would be more productive private investments. However, crowding out doesn’t occur during a BSR because there are no private borrowers to be crowded out. So, strawberry jelly or the knight?


Other things:

The QE Trap is the idea that rates will increase and dampen demand when the thought simply of winding down QE is floated, and that the subsequent recovery following QE is prone to hyperinflation as banks, having access to substantially increased reserves can irresponsibly extend private credit.

The main issue with QE is that the Treasury’s account at the Fed is insufficient to wind down/redeem all the purchased bonds, and thus the Treasury would need to issue new refunding bonds or increase taxes to pay off its debts, both of which have a chilling effect.  

Initially the G20 in 2010, fearing it to be a stealth devaluation tactic, reacted negatively to US QE plans, but the US went ahead with it anyway. 3M later USD depreciated 18% in JPY terms. 

The discontinuation of QE after QE1/2 did not increase rates, implying that the programs themselves had done little to bring down rates. Beyond this, there is a fundamental problem with QE in that it distorts the signaling function of bond yields: low rates imply that there is little private demand/investment and the government should run a deficit to fill the gap, while high rates imply bonds (government borrowing) is crowding out private investment.

The Japanese experience explains the need to promote borrowing and not lending when in a BSR. The prevalence of debt-averseness is a psychological phenomenon and a product of the bitter experience of paying down debt after the 90’s bubble. 

The EU experience is that the lack of economic convergence between core and peripheral countries fueled bubbles in less developed peripheral countries that absorbed excess savings from countries such as Germany and France. 

These bubbles made collapse inevitable as unit labor costs in the EU excluding Germany rose 30%, of which half can be attributed to ECB monetary accommodation, the other half from German labor reforms (p.210)

The subsequent restraints imposed by the Maastricht Treaty’s forbidding of any EU member running a >3% GDP deficit made recovery almost nigh impossible. 

Koo suggests the Eurozone preventing foreigners from buying government debt, and different risk weightings for domestic/foreign bonds to give fiscal space

Distinguishing between a structural recession and a BSR involves investigating the discrepancy between the desires of commercial banks to lend money, and the actual money lent. In a structural recession, banks aren’t willing to lend money, and so no money is lent. However, in a BSR, although banks are willing to lend money, no money is lent as there is no demand for their loans.

Different responses required for 5% and 95% situations: in the 5% crisis situation, lenders need to take losses/haircuts on NPLs which need to be disposed of to keep the financial system healthy and deter agents from taking too much risk; however in the 95% situation, important lenders need to have their liabilities guaranteed by the government and NPLs cannot be allowed to fail (US policy of “pretend and extend” – encouraging mortgage lenders to continue their risky destabilizing practices) to keep the contagion from spreading.