Are Refis Contractionary?
Brad DeLong got me thinking a few days ago, and not in a good way, when he quoted the brilliantly (and brilliantly-named) Cardiff Garcia:
Thus far the surging mortgage origination business at banks has been concentrated in refinancing rather than purchases. The refi boom is great but can only last so long, as Dudley writes, and from a macroeconomic perspective has less of an impact than a housing purchase and construction rebound…
I’m thinking that rather understates the case.
There are at least two of us on this blog who have benefitted from recent refis. (I won’t out the other one, save to say they got better terms than I did.) In my case, the net savings in payments was about $700/month—not exactly chump change, unless you’re Tagg Romney.
But let’s follow the flows here, pretending that all transactions are with two banks for reasons that will be clear. My refinancing means that Bank A receives a lump-sum payout of the balance of my mortgage. But then they have to put that money to work—and they are not going to receive my old interest rate on those replacement loans (if any).
Bank B is receiving current market rates on the refi. So it’s new lending to them. But that’s neutralized on the supply, replaced by Bank A having “freed up” my old loan.
Similarly, on the demand side, my demand is satisfied—and my demand now is $700/month less than it was before.
So the S-D lines are stable for the assets—or even reduced due to the decline in demand. Meanwhile, the flows into financial institutions (assets to them) are reduced.
So if A = L + E and A is reduced, what happens to Liabilities?
If you assume standard economic theory, I save that $700. (Realistically, I spend it and someone else saves it, but the net savings in the economy still goes up $700.) That savings is a Liability, let us say to the Bank A.
Bank A’s Balance Sheet is now: Assets down $700/month, Liabilities up $700/months. To balance that equation, Equity has to go down $1,400/month.
Bank B has Assets up $700/month, Liabilities unchanged, and therefore Equity up $700/month.
Net for the system is that Assets are unchanged, Liabilities are up $700/month and Equity is down $700/month.
Ceteris paribus, refinancing reduces the inefficiencies in the banking system (the above-market asset valuation is replaced by an at-market asset). In doing so, it reduces Bank Equity and increases Bank Liabilities.
Unless lending to other sectors of the economy increases, refinancings that do not take cash (“equity”) out of the property appear to be contractionary.
Ken here is an article on just that from a slightly different perspective.
This brings up my second criticism with regards to the interest income channel. Lowering rates in general in the first instance merely shifts interest income from ‘savers’ to borrowers. And with the federal government a net payer of interest to the economy, lowering rates reduces interest income for the economy.
The only way a rate cut could add to aggregate demand would be if, in aggregate, the propensities to consume of borrowers was higher than savers. But fed studies have shown the propensities are about the same, and, again, so does the actual empirical evidence of the last several years.
And further detail on this interest income channel shows that while income for savers dropped by nearly the full amount of the rate cuts, costs for borrowers haven’t fallen that much, with the difference going to net interest margins of lenders. And with lenders having a near zero propensity to consume from interest income, versus savers who have a much higher propensity to consume, this particular aspect of the institutional structure has caused rate reductions to be a contractionary and deflationary bias.
Even if “fed studies” show the propensity of borrowers to spend to be the same as for lenders, it is almost certainly not so for the borrowed funds. The very act of lending implies saving (not spending) on the part of the lender and the act of borrowing, if it is “non-financial” borrowing, strongly implies spending the borrowed funds, rather than saving.
Once we have that, then there obvious question is, doesn’t this lending transaction represent the marginal consumption choice of both the borrower and the lender? The marginal result is more spending. After all, doesn’t I=S?
That, I think, is where Ken’s analysis may be leaky. “Similarly, on the demand side, my demand is satisfied—and my demand now is $700/month less than it was before.” Nope. You are buying the same house as you were when it cost $700 per month more. In addition, you will spend out of that $700 at your personal MPC, and that will represent new demand from you. If the saver has a lower MPC than you do, there there are more real economic transactions a a result.
Intuitively, I agree with kharris. You now have an additional $700 to spend, or not. But you will spend some of it. Very likely the bank would not.
So, are you arguing that the Fed should be increasing interest rates lest too many people refinance their loans resulting in a weaker economy? Could the Fed have the whole interest rate thing backwards? (They might, you know.) Or … you may personally simply save the $700 a month, but in most cases, the refinancing party would be more likely to spend the money than the bank, high ticket depositor or mortgage note purchaser.