Why Goldman accused the Fed of killing jobs
On April 30, Goldman Sachs sent a letter to the Fed commenting on the implementation of a number of regulations. The Fed then published the letter and the FT quickly zeroed in on the most explosive claim it contained: if counterparty exposure limits were implemented as proposed, real GDP growth would slow by between 0.15 and 0.40% per year. The result, said Goldman, would be the loss of between 150,000 and 300,000 US jobs.
300,000 US jobs, killed by the Fed with a single rule. In an economy struggling to create 115,000 jobs in single month, the idea that regulation could wipe out more than twice that number is alarming. And very specific. As Mark Gongloff noted, almost too specific.
The Fed’s proposed rule limits the amount of exposure financial firms can have to a single other counterparty. Large firms like Goldman would be limited to 10% of total credit exposure, while smaller firms would be limited to 25%. The idea is that if a central counterparty goes bust, it won’t take other big players down with it.
Goldman wants to make the point that the Fed’s proposal will hurt US financial markets and in turn the broader economy. If a bank is underwriting a bond, it will want to hedge that risk by purchasing CDS. The problem, Goldman says, is if the bank can’t buy credit protection from the big sources of such protection, because it’s bumping up against the Fed’s counterparty limits, then it would be forced “to buy either more expensive or less effective hedges” instead. In turn, those excess costs would then be passed along to corporate borrowers.
But how, exactly, did Goldman get to its number?
I did some digging and this is what they did: both the Fed and Goldman use something called the Current Exposure Model, or CEM, to evaluate risk. Goldman took the CEM and adjusted it for how they would change their behavior under the new counterparty exposure limits. Goldman then asked its bond desk to price what the market would look like for corporate bonds under these new (hypothetical), internal risk and funding conditions.
Goldman found that the biggest and most highly rated company yields were unaffected, but smaller companies would get hit with higher yields. These smaller companies aren’t mom and pops, or they wouldn’t raising debt in the capital markets, but they’re also not marquee, AAA corporates. Goldman then took the difference between current yields and what the new model produced, and plugged that difference into the model Goldman’s chief US economist Jan Hatzius uses to predict GDP and job growth.
Shorter version: the outputs from one model were fed into another model whose output was then fed into a third model. Tweaks to the Current Exposure Model fed into a model predicting bond spreads, which were then fed into another model predicting jobs. It’s all quite tenuous, and the error bars are surely enormous.
But what about the larger question of whether higher corporate bond yields are a real job killer? The answer to me seems to be, sort of. This chart shows Moody’s Baa rated yields, in red, which seems like a good proxy for the kinds of companies Goldman thinks might see increased yields. It also shows the unemployment rate, in blue:
Yields and unemployment have moved together since the 60’s, but yields and unemployment have also moved with recessions, particularly in the 70’s and 80’s. Yields are also now extremely low by historical standards, and declining yields are not currently increasing employment.
I’d love to know the specific increase in yields Goldman is predicting that would cause a loss of 150,000-300,000 jobs. But irrespective of the exact relationship their models show between yields and jobs, there is still a great degree of uncertainty, which Goldman is probably willing to admit privately.
If that’s the case, why make the claim? Somewhat perversely, Goldman gave the Fed that number because the Fed didn’t ask for it. Goldman was upset that the the Fed didn’t provide “justification” for its counterparty exposure rule, and didn’t solicit specific estimates of the impact the rule the would have. So Goldman went ahead and came up with their own estimate, dropped it into a letter to the Fed, and the Fed, in a moment of transparency, published the letter on its website. Whether of not this was worth it to Goldman, it’s hard to say.
But the larger claim that limiting counterparty credit exposures in this specific way will hurt the economy gets at two fundamental and related issues: who provides liquidity, and whether liquidity itself will reduce unemployment. Goldman’s claim rests on the belief that liquidity is provided to medium-sized companies by the capital markets. But what this overlooks is that the Fed has been the key provider of liquidity to the capital markets since the financial crisis. And despite that liquidity, unemployment has remained stubbornly high. It seems that the problem is not the supply of debt finance, but rather low borrower demand for those funds. Increasing corporate demand for debt depends on a number of factors, but counterparty exposure limits and a potentially minute increase in rates that are at historic lows is not one of them.