A random walk through the Australian federal election
Plus, reaffirming the Statement on the Conduct of Monetary Policy (for now)
The federal Coalition has a habit of losing elections with the unemployment rate at multi-decade lows. In November 2007, the unemployment rate was 4.4%, on its way to a cycle low of 4% in February 2008 under a new government. We would not see 4% again until February this year.
If that seems unfortunate for the Coalition, it is also unfortunate for the ALP in suggesting that it tends to assume office near business cycle peaks (whether 2022 is such a peak remains to be seen). The popular perception (as suggested by opinion polls) that the Coalition is the better economic manager probably comes from this intersection of the political and the business cycle.
This seems to be a partisan inversion of the stylised US pattern in which the Democrats assume the Presidency at business cycle lows, which then sees the economy outperform on their watch due to a cyclical recovery. That stylised pattern aside, President Biden has also struggled to turn a tight labour market into political support:
While I don’t find partisan models of the business cycle particularly useful, we can’t completely dismiss the possibility of partisan cycles either.
We suggested last week that financial market prices on the Monday after Saturday’s election should be informative of any expected partisan differences in macroeconomic performance. Monday’s mid-morning prices are shown below, benchmarked to other markets:
AUD-USD 0.7089 (+0.7%) (CAD-USD +0.3%)
10-year bond yield 3.28 (-10) (US 10-yr bond yield -9)
June 22 IB futures 99.45 (steady)
ASX200 7163 (+0.24%) (S&P500 +0.57%)
The change in financial market prices does not suggests a significant shift in expectations for the Australian economy.
Simon Jackman shared his betting market-derived time series of the implied probability of winning (IPOW) for a Labor majority government since the election campaign was called. The series passes standard tests for a random walk (technically, we accept the null hypothesis of a unit root process in the Labor IPOW), suggesting that particular betting market was efficiently priced.
We can look at the relationship between the implied probability of a Labor government and financial market time series to gauge whether changes in financial market prices affected Labor’s chance of winning. The idea here is that those changes are proxies for expectations for the economy, which were in turn affected by incoming economic data during the election campaign. Since the election outcome does not appear to have shifted expectations for the economy, we can plausibly treat economic conditions during the campaign as exogenous with respect to the election odds. Among other things, we can think of this as a test of whether there is any disconnect between expectations in betting and financial markets.
In a bivariate model, with financial market variables dynamically pre-determined with respect to the election odds, I find quantitatively small and statistically insignificant effects of the exchange rate and bond yields on the IPOW for a Labor government. Here is what Labor’s IPOW impulse response function (blue line) look like for a one standard deviation shock to the AUD TWI during the campaign period:
A positive TWI shock knocks about 1.5 points off Labor’s IPOW after two days, but the standard error bands (red dashed lines) indicate this is not statistically different from zero. Statistical significance aside, it’s not a big effect. Bond yields look similar. Whatever good economic news the government was getting during the campaign, it was not going to be decisive based on these estimates.
The election campaign is a very narrow window, but we had some big data releases and a change in the official cash rate come in during the campaign. Perhaps what we are seeing is the net of positive economic news like a low unemployment rate and associated expectations for higher inflation and interest rates in terms of their effects on voting intention. We previously made the case for using a quadratic central bank loss function as a model of voter behaviour. On this occasion, the loss on the inflation part of the loss function may have offset the gain on the unemployment/output gap. As I suggested in that previous post, a fully employed economy could still be viewed as disappointing from the standpoint of a voter with expectations derived from trend growth rates in previous decades. The disappointment lies not in actual growth outcomes relative to potential, but potential growth relative to these long-run expectations. Next week’s Q1 Australian GDP release is shaping up as a soft number.
Speaking of efficient markets, RBA Assistant Governor Chris Kent this week followed Governor Lowe in arguing that the effects of QT were already in the price for Australian financial markets:
The speech includes some nice charts of the balance sheet run-off from bond maturities, which peaks in 2025. Interestingly, Kent noted that the Board decided against actively selling its bond holdings on the basis that this would set up expectations for a rapid unwind of future QE programs, undermining their potential effectiveness. That’s a welcome sign that the RBA is thinking about how it might use QE again in future. Some would dispute whether QE can even be effective in the context of efficient markets. I believe it is, but fully accept there are strong arguments going the other way.
This is in contrast to the way in which the Fed’s QT program, which begins on June 1st, is being viewed by markets. As advertised, the Fed’s QT program would imply a balance sheet contraction from 37% to around 20% of GDP by the end of 2024. Market analysts expect the program to contribute the equivalent of anywhere from 25 to 125 basis points of additional tightening, what The Economist euphemistically calls ‘a remarkably wide range.’ The RBA would have us believe that all that matters from here on out is the change in the cash rate, with QT already in the price, but markets view the Fed’s QT program as giving rise to much more uncertainty and as a potential substitute for prospective increases in official interest rates.
The RBA has also released a RDP on the effect of its bond purchases to support market liquidity, its 3-yr bond yield target and its quantitative easing program. The results are broadly consistent with the NY Fed study we linked to earlier in pointing to fairly narrow liquidity effects. In my view, this just makes the case for future programs being broader in scope, with a wider range of eligible securities.
John Kehoe had a story confirming that the existing Statement on the Conduct of Monetary Policy will remain in place pending a post-election review of the RBA and monetary policy. Kehoe notes the various models that are being considered for the review, although my understanding is that the new Labor government is leaning towards having an overseas expert preside over the review process.
One focus for a review is likely to be governance and accountability mechanisms. As Myriam Robin notes in The Curse of the House Economics Committee, all but one of the Coalition members of the Committee lost their seat at the election. The Committee is an imperfect accountability mechanism at the best of times and is hardly going to be strengthened with the loss of nearly every incumbent Coalition member.
ICYMI