The Covid-19 shock almost caused a financial crisis. Central banks reacted aggressively to stop the free fall. My research with Ricardo Caballero sheds light on these events. Long thread! 1/N #EconTwitter
VoxEU tinyurl.com/y9wd9g98
Paper tinyurl.com/ybzl3wkn
VoxEU tinyurl.com/y9wd9g98
Paper tinyurl.com/ybzl3wkn
The economy produces goods and services that must be absorbed by spending by households, firms, governments... Much of modern macro is concerned with equilibrium in goods markets 2/N
Economic activity also involves risks that need to be absorbed by investors---banks, institutions, households... Much of modern finance/asset pricing is concerned with equilibrium in risk markets 3/N
Risk markets feature substantial heterogeneity. Expert financial institutions value risks very differently than households. Optimists might have a different view on house price risks than pessimists. Recent financial crises showed the importance of heterogeneity 4/N
Goods and risk markets are usually analyzed separately but they are related. Asset prices provide the missing link. As we will see, they affect demand in both markets. So they provide a link by which problems can spill from one market to the other 5/N
Ricardo Caballero and I have developed a framework that captures the links between goods and risk markets. We have used it to study a number of issues. Most recently, we applied our model to the Covid-19 episode 6/N
Most economists view Covid as a supply shock: a decline in productivity. In our model, this spills to risk markets, damages risk-tolerant investors, and spills back to the goods market through demand. A supply shock can lead to a severe demand recession! Let's go over details 7/N
Start with the goods market. Much of the empirical evidence shows asset prices affect demand. Credit spreads reduce firms' investment and households' spending. House prices affect spending--see the excellent work by @AtifRMian and @profsufi. Stock prices also matter 8/N
A key block of our framework is the output-asset price relation. A stylized equation that captures numerous linkages between asset prices and demand. We define also the asset price per supply: this helps to think about a supply shock (Covid) 9/N
Given the increasing relationship between asset prices and demand, there is an asset price level consistent with efficient output. For a healthy economy, asset prices (broadly construed) should be in proportion to supply 10/N
When asset prices are too low, demand is low and this triggers a recession. Supply shocks can turn into a demand recession when accompanied by large declines in asset prices! 11/N
Asset prices are determined in risk markets. A key determinant of risk demand is the Sharpe ratio: risk premium normalized by risk. I will treat Sharpe = risk premium (distinction not important for this thread) 12/N
Sharpe ratio depends on expected return above risk-free rate. A lower asset price or risk-free rate raises the Sharpe for risky assets. This increases demand for risk. Notice how the interest rate policy made an entrance: this is how the Fed affects the economy in our model! 13/N
The most important block of our model is the risk balance condition. It formalizes these effects and describes the asset price that ensures equilibrium in risk markets. A key parameter shows up: effective risk tolerance... 14/N
Effective risk tolerance controls the appetite for risk. With lower tolerance, investors require a higher Sharpe ratio (or risk premium) to hold the risk supply. This is how investor heterogeneity enters our analysis... 15/N
With heterogeneity, effective risk tolerance depends on the wealth share controlled by risk-tolerant investors. We loosely refer to them as "banks" but have much broader interpretation in mind. Many financial institutions fall into this group 16/N
"Banks" are more exposed to risky asset prices than "households." In our model, banks take on leveraged risks (and households don't). This is obviously stylized: captures the many ways in which risk-tolerant are naturally exposed more to economic risks 17/N
With leverage, a surprise decline in asset prices creates disproportionate damage to "banks'" wealth. This is why the financial crisis was so severe. This time, actual banks had less leverage but the shock was larger and less predictable 18/N
This yields our third equation that links risk tolerance to asset prices. A surprise supply shock reduces risk-tolerant wealth more than risk-intolerant wealth, pulling down the average risk tolerance 19/N
Putting everything together, we have our three equation model. Given the risk-free rate, everything can be solved. Risk-free rate is the policy variable. We assume the Fed tries to set it to ensure the efficient asset price and output. But it might be constrained (or late) 20/N
This is obviously a highly stylized model and misses a lot of richness. But it helps to capture the general equilibrium interactions between goods and risk markets--which many other models miss. Think IS-LM with risk markets 21/N
Let's apply our model to think about the Covid crisis. We capture this as a large and (hopefully) temporary shock to supply/productivity. This naturally lowers asset prices and output. But not price per supply. Prices should fall as much as supply but shouldn't overshoot 22/N
Since asset prices fall but future payoffs are unchanged (assuming the shock is temporary) the expected return increases. If this was the only effect, then the Fed would have to raise the risk-free rate to ensure asset prices are not too high! 23/N
Intuitively, markets are forward looking. If the shock will pass soon, and risk tolerance is unchanged, then risky assets would be attractive. Price per supply would rise and demand wouldn't fall as much as supply. The Fed would have to raise the rate to ensure demand=supply 24/N
But a low asset price also reduces the risk tolerance. The required Sharpe ratio increases. This swamps the expected recovery effect for a large shock or high leverage. Market requires a greater Sharpe to hold the same risk. The Fed needs to cut rates to raise the Sharpe! 25/N
The risk-tolerance effect also dominates (regardless of size of shock or leverage) when the market thinks the shock is more persistent--a distinct possibility. When there is no recovery in horizon, there is no expected-recovery effect! 26/N
This can easily push the interest rate into the zero (or effective) lower bound. This is what happened in much of the developed world. Many emerging market monetary policymakers are also effectively constrained: lowering rates can trigger a large currency depreciation 27/N
When the Fed cannot cut the interest rate sufficiently, the decline in risk tolerance causes asset prices to decline. Investors get their high required Sharpe one way or another. But adjustment of the risk balance via asset prices is costly, for two reasons 28/N
The excessive decline in asset prices causes a demand recession: the output falls more than the supply. A "real" shock that originates on the non-financial side spills to financial markets. Then it spills back to the goods market via demand and exacerbates the recession! 29/N
Importantly, the excessive asset price decline also triggers a downward loop. Low asset prices amplify the damage on the risk-tolerant investors, which further lowers the risk tolerance. This raises the required Sharpe ratio more, which lowers asset prices more, and so on 30/N
In practice, many frictions such as bank runs or illiquidity would make the downward spiral more severe. But our model has a downward loop without those frictions! This means traditional policies such as the lender of last resort will help but won't fully solve the problem 31/N
So what policies work then? Policies where the Fed (with the backing of treasury) absorbs some risk can be very effective and reverse the free fall. We call these policies large-scale asset purchases (LSAPs) but have in mind many other recent facilities with similar flavor 32/N
LSAPs work by transferring risk to the government's balance sheet. This reduces the required Sharpe ratio: with less risk, the market requires less compensation. This counters the effects of reduced risk tolerance: it stabilizes asset prices and reverses the downward spiral! 33/N
LSAPs are not a handout to "banks." They benefit banks indirectly (and this is why they are powerful) but they also raise output and improve everyone's income and wealth--including the less risk tolerant. Without LSAPs, there would be a worse recession and a financial crisis 34/N
LSAPs don't cost much either. Government buys assets at a low price and a high Sharpe ratio. Taxpayers on average gain! But LSAPs are not free: taxpayers lose money in extreme bad states. So LSAPs require fiscal capacity: ability to raise taxes or cut spending if necessary 35/N
So what is the policy message? The CBs should be applauded for aggressive interventions that prevented a crisis. They went beyond the usual lender of last resort and became the "investor of last resort." The danger is not over and they should be ready to do more if necessary N/N
