ABCT and the Community Reinvestment Act

Austrian business cycle theory explains the general pattern of the boom-bust cycle — credit expansion, lowered interest rates, malinvestment, crash, liquidation — but the particulars differ in each historical case. (Austrians sometimes distinguish “typical” from “unique” features of each cycle.) To explain particular episodes, we appeal to specific technological, regulatory, political, legal, or other conditions. For example, in the 1990s, much of the malinvestment was channeled into the IT sector, where uncertainty driven by rapid technological change made entrepreneurs particularly susceptible to forecasting errors. In the 2000s, of course, malinvestment appeared largely in real estate, the result of government programs designed to relax underwriting standards and otherwise increase investment in particularly risky real-estate assets. In other words, ABCT tells us to look for malinvestment during the boom, but not where that malinvestment will show up.

Regarding the latter example, however, there has been a persistent dispute among mainstream economists about the role of government housing policy, particularly the Community Reinvestment Act which was used, in the 1990s, to make banks increase their lending to particular low-income neighborhoods. Paul Krugman asserts, for example, that the “Community Reinvestment Act of 1977 was irrelevant to the subprime boom.” Actually, no. A new NBER paper (gated) on the CRA is causing quite a stir. Authored by four economists from NYU, MIT, Northwestern, and Chicago, the paper is the first to use instrumental-variables regression to distinguish changes in bank lending caused by the CRA from changes that would likely have happened anyway. (The authors use the timing of loan decisions relative to the dates of CRA audits to identify the effect of the CRA on lending.) The results suggest that CRA enforcement did, contra Krugman, lead banks to make substantially riskier loans than otherwise. Raghu Rajan puts it in a very Austrian-sounding way:

The key then to understanding the recent crisis is to see why markets offered inordinate rewards for poor and risky decisions. Irrational exuberance played a part, but perhaps more important were the political forces distorting the markets. The tsunami of money directed by a US Congress, worried about growing income inequality, towards expanding low income housing, joined with the flood of foreign capital inflows to remove any discipline on home loans. And the willingness of the Fed to stay on hold until jobs came back, and indeed to infuse plentiful liquidity if ever the system got into trouble, eliminated any perceived cost to having an illiquid balance sheet.

I’d reverse the order of emphasis — credit expansion first, housing policy second — but Rajan is right that government intervention gets the blame all around.


  1. A few things usually missing:

    Home price inflation* above their value — unsustainable/could not be maintained w/sufficient new buyers — contingent upon ever rising prices. Look at the price-income ratio; normally a person buys a home priced at 3-4 times their income, but there were many mortgages of higher magnitudes.

    So, the collateral (the home) to resell on default & foreclosure needs to maintain the high price for near break-even.

    *The inflated home prices resulted from the simple supply & demand principle, but few know the details & reasons.
    Nation-wide, there was rather uniform increased demand from lower interest rates & loose lending standards, but prices vastly increased in some metropolitan areas (mostly coastal & 2 desert) — those having excessive restrictions/zoning — much land is off limits or takes time & money to allow for building (ie SF Bay Area).

    The geography of actual physical land shortage (ie Manhattan & Honolulu) is minor. The home supply not keeping pace with demand is measurable — the occupancy rate is lower in those markets, usually <half of national — was ~9%, became ~11%.

    • The idea that banks just, for no reason, allowed “more leverage thus more risk” to enter the market ignores the basic math.

      Debt/income ratios increased because the rates went rapidly down, thus dramatically and suddenly increasing the amount of debt that could be serviced with the same income.

      A $350K ARM priced in 2003 at a three-year rate of 3.75% on a 30 year amortizing schedule, will have payments of $19,451/year. Two years earlier, at 5.75%, the payments would have been $24,510/year. That is a change from $2,042/month to $1,621/month. A household with $25K free cash flow would not qualify for the loan at 5.75% but would qualify at 3.75%, with a 1.29x debt-service-coverage ratio. 0.29x is not super but would typically be viewed as adequate cushion – especially for people with good credit, as it would be assumed that they would cut back elsewhere to make payments in a rising rate environment, and/or refinance on a fixed rate, and/or sell into a rising-price environment.

      That same household would see increasing property taxes and food and energy prices over the next three years, thus even if someone got a raise, the net free cash flow would probably not change much.

      Then the loan re-priced in 2006 at 6.25%, for annual debt service of $25,860, monthly debt service of $2,155. Payments are up by $534/month and the homeowners can no longer make them. They might put the house on the market but so do others in the same situation, and the houses do not sell.

      There would have been a bubble, bust and recession without the “relaxed standards” lending, most of which took place after the boom got going.

  2. “So risky mortgage lending didn’t cause the crisis. What (partially) caused the crisis was risky mortgage lending being mistaken for non-risky mortgage lending, by people who ought to have known better.”

    What should they have known better? The implications of unorthodox Fed policy. I agree they should have known better, and not everyone does, even now – but between blaming the policy itself and blaming people’s reactions to the policy, I blame the policy.

    But what was risky about the mortgage lending? What’s the risk? That the ARMs and HELOCs would re-price, and that with higher rates no longer driving asset prices up and creating new spendable cash, purchases would fall and unemployment would increase. What’s the risk? Leverage? Why was there more leverage? Primarily because when rates are low, the same income services more debt.

    What was risky about the securities and portfolio analysis? That historical delinquency and default rates were used to predict future tendencies? Well when rates fall, so do delinquency and default rates, because you’re lowering the bar. Again – what is it that the analysts failed to predict? Primarily the raising of the bar, and secondarily the impact that all that new money resulting from cash-out refis and helocs was temporarily having on spending and employment as well as home investment and asset prices (i.e., the ability to sell the house when you lost your job).

    In short, the non-Austrians claim “it wasn’t the low rates” but then “it was the risk [of the rates].” That’s circular logic – it does not add up.

    • You typed (paraphrasing) the “claim ‘it wasn’t the low rates, but the risk’ That’s circular logic.”
      That is a misunderstanding, rather than a non-Austrian view. In fact, it’s even another way of looking at “misallocation of resources.”

      I don’t see you view that explanation as a tautology. Interest rates reduces the eventual loan cost, which is independent of whatever the loan is used for & the collateral.

      Some mortgages were riskier for several reasons.
      1.Amount of down payment
      2.Verifiable, sufficient & longevity of borrower’s income to maintain mortgage
      3.Resellable home at purchase price or mortgage balance

      • The riskier loans you reference (e.g., 0% down, liar-loans) were for the most part late in the game – after the peak, well after the banks had built up lending infrastructures, as rates had started to come back up and rate-driven demand for refinancing volume had fallen off. There would be no such need to maintain volume for huge lending infrastructures if the huge lending infrastructures hadn’t been built up, which wouldn’t have occurred without the initial bubble, which was caused by the rates. These are high risk but were seen as lower risk than they actually were because the housing values had gone up 10% / year, which was the result of the lower-rates allowing people to take on more debt using conventional ARM structures.

        The notion that everything was just fine until a lot of banks for no apparent reason decided to do no-doc, no money down loans, is not accurate. Those decisions were made in response to a series of events that wouldn’t have occurred but-for the low rates.

  3. I don’t understand how you get from point A to point B here. Yes, the working paper demonstrates that the CRA lead to riskier lending. But how does that suggest at all that Krugman was wrong that the CRA was irrelevant to the crisis?

    Surely you’re not suggesting that riskier loan portfolios are predicted by ABCT (or any!) theory to result in financial crises, are you?

    Noah Smith put it well here:

    He writes: “So risky mortgage lending didn’t cause the crisis. What (partially) caused the crisis was risky mortgage lending being mistaken for non-risky mortgage lending, by people who ought to have known better.”

    I think the evidence suggests that Agarwal et al. are right and Krugman is right, and I don’t think those two things are as contradictory as you’re suggesting here.

Leave a Reply