Consumer Credit: Too Much or Too Little (or Just Right?)

Consumer Credit:
Too Much or Too Little
(or Just Right?)
Jonathan Zinman
Dartmouth, IPA, NBER, etc.
October 2013
• Many policy efforts to increase consumer credit
– Loan guarantees (e.g., FHA)
– Public provision/subsidies (e.g., student loans)
– Concern about gaps in the “lending ladder”
• Many policy efforts to (functionally) decrease
consumer credit supply
– Interest rate and fee ceilings (consumer loans, credit
– Restrictions on rollovers (payday loans)
• These policies presume answer to threshold questions:
– “Is credit under-supplied?”
– “Is credit over-supplied?”
• If “yes” to either, or both, we’d also like to know:
– “When, where, and why?”
– (What’s the failure: market, decision, and/or policy?)
• Evidence on these questions would give us better
sense of which policy options likely to:
– Be (constrained) optimal
– Pass cost-benefit test
• And hence which options worth testing, evaluating,
simulating, etc.
Is Credit Under- or Over-Supplied?
• Paper reviews what we actually know about these
questions, theoretically and empirically
• Organized around 8 classes of theories of inefficient supply
• For each of these theories, consider empirical evidence:
– From U.S. (mostly), and big markets therein: mortgage, auto,
student, credit card, payday, bank overdraft
– That convincingly tests theory
– That is more “circumstantial” (consistent with many theories),
yet oft-invoked as supporting a particular one/class
• Also review reduced-form evidence on “impacts”:
downstream effects of exogenous changes in credit supply
on consumers
Plan for Today
• Briefly sketch the 8 classes of theories
• One sentence summary for each of related
empirical evidence
• Summarize key takeaways
Theories of Under-Supply:
1. Market Power
a. Plain vanilla
– Pretty implausible for most part in U.S.
b. Search/switch costs
– Empirical evidence consistent with very high costs
– Implications for efficiency depends on other
Theories of Under-Supply:
2. Regulatory Failure
Motivating facts:
• gaps in “lending ladder” between most expensive
credit card & cheapest payday loan
• Many state laws outlaw the very contracts that
would fill these gaps
Tests of this theory?
• None that I know of
• Gaps could also be consistent with asymmetric
Theories of Under-Supply:
3. Asymmetric Information
a. Ex-ante hidden information: adverse selection on
type, project risk
b. Ex-post hidden action: moral hazard in project choice,
effort broadly defined, strategic default
c. Selection on moral hazard
Taking theories to the data:
• Doable, but still limited quantity of evidence
– Karlan and Zinman (and Bryan)
– Einav and Levin and co-authors
• Mixed on which varieties matter
Theories of Over-Supply:
1. Asymmetric Information
a) Advantageous selection
b) Moral hazard with common agency/common
pool/coordination problems
Taking theories to the data:
b) Seems quite plausible, but consumer credit evidence
quite thin
a) Seems somewhat plausible, given:
• Evidence from insurance markets
• Marginal borrowers default more when subject to
higher downpayment in Einav et al (ECMA)
Theories of “Over”-Supply:
2. Borrowing with Negative Externalities on Collateral
a) Fire sales
b) Location, location, location (3L) effects in real estate
Taking theories to the data:
• Limited
• Mixed (Mian and Sufi infer fire sale but no 3L,
Campbell et al v. versa)
Theories of “Over”-Supply:
3. Deleveraging Slows Recovery
• If leverage high enough entering shock, then the
larger the debt overhang, the larger the rate cut
needed for output to stay at potential. Zero
bound may bind (Eggertsson and Krugman 2012;
also Hall 2011)
• Taking theories to the data: Mian and Sufi (2011,
2012, 2013) find that drop in demand more
pronounced in counties and zips that entered
downturn with relatively high leverage
Theories of “Over”-Supply:
4. Systemic Risk via “Ratchet Effect”
Khandani et al (2012): unintentional synchronization of
homeowner leverage from
• rising home prices
• declining interest rates
• friction-free refis– and asymmetry therein (can ratchet
leverage up but not down)
Taking theory to the data:
• Not much. Bhutta and Keys (2013) find that later
extractors more likely to default than early extractors
Theories of Over-Supply:
5. Behavioral Consumers
Different varieties. Most interesting ones posit a
plausible bias re:
a. Preferences
b. Expectations (over-optimism)
c. Attention
d. Price perceptions
Taking theories/hypotheses to the data:
• Limited but intriguing evidence
– Except for over-optimism, which is 0 for 2 so far
Reduced-Form Evidence on the
“Impacts” Question
• Do proxies for consumer welfare get better or
worse as credit supply increases or decreases?
• Evidence quite mixed, especially if restrict to U.S.
• Even a consensus finding would be difficult to
map into theory and policy
Policy/enforcement asymmetries
Heterogeneity => targeting and screening?
Credit market failures*other market failures?
Key Takeaways: Conceptual
• Theories abound for both over-, under-supply
• Several classes have elements of both:
– credit over-supplied heading into a downturn
– then under-supplied during the downturn
• If these models correct, then which objective
should policymakers adopt in calmer times?
– Try to reduce frictions, in anticipation of next
– Or assume that frictions will always be severe enough
to exacerbate downturns, and introduce frictions to
keep leverage form being too high at next downturn?
Key Takeaways: Empirical
• Most classes of theory have some support
• No class of theory has anything resembling
definitive/overwhelming/consensus support
• Overall lack of evidence on:
– Whether markets err
– In which direction
– When/where/why
• Humility! Traditional approaches to policymaking
will likely do more harm than good in many cases
Key Takeaways: Methodological
• Allow for heterogeneity: seems likely that
diagnoses and treatments will vary substantially
across product markets
– Attract different (behavioral) types of borrowers
• Interactions between type and opportunity sets may be
– Contracting equilibrium merits scrutiny. One e.g.:
• More sympathetic to concerns about a market with strong
symptoms of shrouded pricing (bank overdrafts) than one
where pricing more-transparent (payday)
– Collateralized vs. not likely important. Two e.g.:
• Externalities on collateral assets
• Bankruptcy rules
Key Takeaways: Methodological
• So what should policymakers do?
– I am not advocating a simple “hurry up and wait for
more research” strategy
• Pilot-testing new regulations and policies, in ways
that generate rigorous evidence
– More difficult for stabilization policy, more feasible for
rulemaking focused on steady-stage
– Safe harbors for firms that experiment on policyrelevant margins
• I.e., more “beta-testing”, less “ready, fire, aim!”

similar documents