Australians more easily stimulated than Americans

I have a new paper out today, looking at the impact of the Australian 2009 household payments on expenditure. It uses a poll question contained in the June ‘ANU Poll’. Here’s the abstract (click on the title for the full paper):

How Much Did the 2009 Fiscal Stimulus Boost Spending? Evidence from a Household Survey
Andrew Leigh
Using survey evidence, I estimate the impact of a $12 billion package of household payments delivered in Australia between March and May 2009. Forty percent of households who said that they received the payment reported having spent it. This is approximately twice the spending rate that has been recorded in surveys assessing the 2001 and 2008 tax rebates in the United States. Using an approach for converting spending rates into an aggregate marginal propensity to consume (MPC), this is consistent with an aggregate MPC of 0.41-0.42. Since this estimate is based only on first-quarter spending, it may be an underestimate of the longer-run impact of the package on consumer expenditure.

Surveys aren’t the ideal way of estimating expenditure, but in my view they’re superior to looking for sudden movements in quarterly macro figures, which has been much of the debate so far.

Advertisements
This entry was posted in Macroeconomics. Bookmark the permalink.

5 Responses to Australians more easily stimulated than Americans

  1. Simon says:

    I think this is a really interesting paper, and a really valuable alternative way of thinking about the problem.

    However, I can’t help but feel that there’s a 500-pound gorilla in the room, in the form of the 32% response rate. Unless we know (or assume) something about (1) the representativeness of the respondents chosen for the telephone survey and (2) the 68% who didn’t respond to the survey, then all of the consequent estimates are unidentified. For me, this should be the most likely explanation for the discrepancy noted on page 6.

  2. Andrew Leigh says:

    Simon, I’m glad you raised this. I published the 32% response rate figure because I think it’s good science to do so. I entirely agree with you that there could be bias here, but it’s unclear how we would check this. (In the old days, phone numbers corresponded to geographic areas, so you could at least do breakdowns by postcode income, but I think that link has now been basically broken.)

    Also, don’t assume that my response rate is lower than other surveys. There is talk in the political polling literature of response rates below 25%. So far as I could tell, none of the other studies that I discuss reported their response rate.

  3. Simon says:

    Yes, I agree — I think things like response rates are not published enough, and I absolutely agree with your having made it clear. I suppose, as you say, these are inherent problems with telephone polls, and problems that we probably just have to live with. I guess my point reduces to the objection that we need to assume that the 32% is a representative sample of the Australian population in terms of its propensity to spend — not a crazy or unreasonable assumption, I think, but a relatively strong one nonetheless, and a further potential explanation for the discrepancy noted on page 6.

  4. simon mongey says:

    Hi Andrew,

    I don’t know if you recall receiving the small survey made using googleforms on this same issue that I sent around that I a while ago? Turns out snowballing was a bad thing! Valuable lesson learnt!

    Moving away from the main aim of the paper, (re: MPC out of the payments) I was very surprised by a couple of things in Table 1. The most prominent being the very small percentage of people that used the payments to pay off credit card debt.

    Assuming that credit card debt levels are at the published levels (~$3000 per person), then its strange to see more people using this influx of disposable income to pay off their lower rate mortgage than their credit card. (Of course this makes an assumption about the distribution of this credit card debt).

    However, even considering this, the paying off of bills over credit cards is also slightly unusual. I would imagine that a large amount of credit card debt is derived from the purchase of utilities. Not only is payment via credit card far more convenient, it is a justifiable use of credit, with the high rate of interest being a fair price for the maintenance of key utilities. Therefore if we see more people paying off current / outstanding bills rather than the credit card, does that then suggest that people have maxed out their cards with previous bill payments?

    Still, that 1.5% is strange, any ideas as to why that would be so low? I agree that 32% is reasonable in a poll for conclusions to be drawn, however could this response rate be at all linked to this number? Maybe, if credit card debt is seen with a sense of stigma, then respondents who have done this will be bias towards choosing paying off bills, or not respond?

  5. Pingback: Stimulus, Again at catallaxyfiles

Comments are closed.