A statistical model of voting intentions continues to suggest that the Conservatives will fail to achieve an overall majority at the 2015 election despite an improving economy.
The model is designed to predict the poll differential between the main government and opposition parties based on economic factors. It was estimated on ICM-Guardian poll data extending back to 1984. The poll differential depends positively on average earnings growth and house price inflation, and negatively on the unemployment rate, retail price inflation and interest rates (Bank rate). More details are available in a previous post.
The Conservatives were 1 percentage point (pp) behind Labour in the June ICM-Guardian poll. The model estimates a 5 pp shortfall based on current economic readings. The 4 pp difference is within the historical margin of error of the model. Other recent polls have been less favourable for the Tories: the average calculated by the UK Polling Report website shows them lagging by 3 pp.
The model estimate of the Conservative / Labour differential has risen from a low of -10 pp last year as the economy has improved. The continued shortfall is attributable to low average earnings growth. According to the model, a big rise in earnings growth would be needed to generate a Tory lead of 6 pp, which is probably the minimum required for the party to achieve a Commons majority*. This is possible but would bring forward Bank rate hikes, with an offsetting poll effect.
The chart, an update from the previous post, shows three possible scenarios. The first is based on the latest MPC economic forecast in the May Inflation Report. In this forecast, average earnings growth rises to 2.75-3% by May 2015, the unemployment rate drops to about 6%, inflation is stable and Bank rate increases by 25 bp early next year. The model predicts a tiny Conservative lead by election time, consistent with Labour obtaining most seats in a hung parliament.
The second scenario is intended to be a “best case” for the Tories. Earnings growth rises to 4% while unemployment drops to 5.5-5.75%, but lower-than-expected inflation results in the MPC keeping Bank rate on hold. The Conservative lead reaches 6 pp by May 2015 – just sufficient to deliver a seat majority.
The MPC’s guidance, however, implies that it would raise Bank rate earlier in this scenario because of a faster erosion of slack. The third scenario uses the same economic assumptions but incorporates 0.25 pp rate hikes in November, February and May (just after election day). These hikes cap the Tory lead at 3-4 pp, suggesting similar numbers of seats for the two parties. Such an outcome might force a repeat election later next year or in 2016.
Non-economic influences could be of greater importance in the coming election, with additional uncertainty injected by the Scottish and promised EU referenda. A stronger economy should continue to aid the Conservatives but is unlikely to be sufficient to deliver victory.
*Each 1 pp rise in earnings growth boosts the Conservative / Labour differential by 4 pp, according to the model. See the previous post for a discussion of the relationship between voting intentions and seat allocations.
Investors are positively inclined towards the dollar. A net 58% of global fund managers believe that the US currency is undervalued, the second highest reading in more than 10 years, according to Merrill Lynch. US futures market investors, excluding “commercials”, are moderately long the dollar against other developed market currencies, based on the weekly commitment of traders report*.
Dollar bulls expect the currency to be boosted by a continued rise in US / foreign interest rate differentials as the US economy grows strongly over the remainder of 2014. Monetary trends support optimism about US economic prospects: six-month real narrow money expansion rose to a 17-month high in May and is stronger than in most other developed economies.
The bullish argument, however, ignores recent deterioration in the capital account of the balance of payments. The current account deficit has been stable but the balance of direct and portfolio investment flows has moved from a significant surplus in 2012 to a small deficit in the latest 12 months – see chart. The “basic balance”** deficit, therefore, has risen to its highest since 2009.
The capital account has been weakened by a rise in portfolio investment outflows, partly reflecting the ebbing of the Eurozone crisis and optimism about “Abenomics” in Japan. Higher US portfolio investment overseas could be sustained as the global economy improves. The direct investment deficit, meanwhile, could widen as US corporations step up foreign take-over activity, partly for tax reasons.
A basic balance deficit requires an offsetting surplus on the short-term capital account. US economic strength and a rise in US / foreign interest rate differentials may attract a larger inflow of short-term capital but there is no guarantee that this will be sufficient to outweigh the wider basic balance deficit, resulting in upward pressure on the dollar. Bulls may continue to be disappointed.
*The net long position against six other developed market currencies was equivalent to 15% of open interest as of last Tuesday.
**Basic balance = current account plus long-term capital flows (conventionally defined as direct / portfolio flows).
Why did UK economic growth take off from late 2012? Economists interviewed recently by the Financial Times variously cited QE, housing market stimulus measures and a spontaneous revival in confidence. These explanations are unconvincing. QE2 started much earlier – October 2011 – and seems to have had a limited impact on monetary trends. The Help to Buy guarantee scheme, meanwhile, came into effect well after the growth pick-up. Confidence did revive from late 2012 but why?
Two factors are judged here to have been key for the economic reversal. First, inflation fell sharply between 2011 and 2012, moving below the rate of broad money expansion. The “real balance effect” – the impact of changes in real money holdings on spending intentions – turned from negative to positive.
Secondly, banks’ wholesale funding costs declined from mid-2012, reflecting improved Eurozone financial conditions and the funding for lending scheme, resulting in a large reduction in retail deposit rates in late 2012 and early 2013. This fall encouraged households and firms to spend more of their monetary savings – the velocity of circulation of broad money, in other words, rose.
The stimulatory effect of lower deposit rates on the economy was signalled by a strong pick-up in narrow money from mid-2012 as funds were shifted from time deposits (“savings money”) to sight deposits (“spending money”).
A representative interest rate on longer-term time deposits has fallen by 1.6 percentage points (pp) since mid-2012 – see chart. This is equivalent to half of the decline between 2007 and 2009, when the MPC cut Bank rate by a total 5.25 pp.
A correct explanation of why growth rebounded is important for assessing current economic prospects and appropriate policy. If QE2 was the key factor, economic momentum should already be fading, since the programme was ended long ago*. The housing stimulus explanation, meanwhile, suggests that the onus is on the government to abolish Help to Buy, rather than on the MPC to raise interest rates. The view that confidence drives economic shifts would also warrant caution about tightening monetary policy, since current “animal spirits” may be fragile.
The “monetarist” explanation offered above, by contrast, suggests that growth will remain strong and may be insensitive to a rise in interest rates. Real broad money expansion is stable and bank deposit rates have fallen further since end-2013. With banks funding comfortably, deposit rates may rise by much less than any increase in Bank rate. Just as earlier Bank rate cuts failed to stimulate the economy because the “transmission mechanism” was broken, a small rise may be ineffective in slowing the economy now that the monetary plumbing has been repaired.
Japan’s QE programme has failed to boost money supply growth, despite its massive scale. This supports US / UK evidence that the monetary impact of QE has been greatly exaggerated by its supporters. The weak monetary response implies that QE has had little influence on economic developments in the three countries. Its main effect has been to encourage speculative behaviour in financial markets, increasing the risk of asset price bubbles.
Incoming Bank of Japan (BoJ) Governor Kuroda significantly expanded its QE bond-buying operation in April 2014. Annual growth of broad money M3 stood at 2.5% in March 2014; some QE optimists expected the new blitz to drive it up to 10%. There were initial indications of a modest positive effect, with annual M3 expansion reaching 3.5% in November. It has since fallen back to 2.6%, as of May.
Why has the liquidity created by the BoJ to buy securities failed to boost the money supply? An answer to this question requires an analysis of the “monetary counterparts”. The “money supply” is the main liability item on the banking system’s balance sheet. Since the system’s total assets equal liabilities, money supply changes can be “explained” by movements in other balance sheet items. The chart shows the contribution of these other items to annual broad money growth.
QE has had the expected expansionary impact on banking system assets: the contribution of the BoJ’s lending to the government to annual broad money growth rose from 3.2 to 5.9 percentage points (pp) between March 2013 and April 2014* (blue bars). This increase, however, has been neutralised by larger sales of JGBs by banks: a reduction in their lending to the government subtracted 2.8 pp from money growth in April, up from 0.9 pp in March last year (red bars). Banks are selling mainly because QE has boosted their reserves at the BoJ, so they need to hold fewer JGBs to meet their liquidity targets.
There has been an additional small drag on broad money growth from a slowdown in bank lending to other Japanese residents: such lending contributed 0.9 pp to money growth in April versus 1.3 pp in March 2013.
To a first approximation, therefore, QE has amounted to a large-scale swap of JGBs and reserves between the BoJ and other banks, with no impact on the money supply. The increase in banks’ reserves, moreover, has not resulted in an expansion of their lending to the rest of the economy.
Remarkably, the joint contribution of BoJ and bank lending to the government to broad money growth is smaller now than when QE was launched in October 2010 – 3.3 pp versus 3.1 pp (sum of blue and red bars). Banks were then driving the expansion, buying JGBs to boost their liquidity ratios.
The BoJ is committed to maintaining securities purchases at their current pace but the monetary impact should continue to be neutralised by bank selling. The stock of bank lending to the government still amounts to 25% of the broad money supply so the recent rate of reduction could be sustained for many years.
The BoJ could try to achieve more bang for its buck by buying equities rather than JGBs. Banks, however, would probably still sell JGBs as their reserves expanded so there is no guarantee that the money supply impact would be larger. Higher equity prices, in isolation, would be unlikely to provide even a short-term boost to the economy, given the absence of a “wealth effect” in Japan.
Rather than BoJ intervention, any rise in broad money growth is likely to be driven by a pick-up in bank lending to the private sector or a stronger balance of payments position**. “Abenomics”, however, has so far failed to stimulate credit demand, while the balance of payments has deteriorated recently, reflecting both a lower current account surplus and net direct / portfolio investment outflows.
Current monetary trends suggest modest economic growth and low inflation – probably beneath the BoJ’s 2% target***. Such a scenario is not unsatisfactory given Japan’s demography and could have been achieved without QE.
*The counterparts analysis is not yet available for May.
**A basic balance (i.e. current account plus net direct / portfolio flows) surplus results in a rise in the banking system’s net foreign assets, included under “other counterparts”.
***Stripping out sales tax effects.
The MPC supposedly places significant weight on an estimate of the “medium-term equilibrium unemployment rate” in calibrating its policy stance. The August 2013 Inflation Report, for example, stated that “The gap between actual unemployment and the medium-term equilibrium unemployment rate is a measure of effective slack in the labour market, and is likely to be most relevant for assessing wage pressures over the MPC’s three-year forecast period.”
So what is the Committee’s current judgement about this crucial policy metric? The May Inflation Report, on close inspection, is contradictory. The text at the top of page 30 of the Report refers to a current estimate of 6-6.5%. Chart 3.7 on page 28, however, contains a first-quarter figure for the “unemployment gap” – the difference between the actual and medium-term equilibrium rates – of 0.9%. Since the MPC expected actual unemployment to be 6.7% in the first quarter, this implies an equilibrium rate of only 5.8%.
How has this contradiction arisen? One possibility is that Bank of England staff generated the 5.8% estimate but the MPC majority refused to endorse their assessment, preferring a 6-6.5% range. Chart 3.7 may have been retained in the Report owing to an editorial insight, or because the still-significant unemployment gap shown supports the dovish policy stance of Governor Carney.
The single-month unemployment rate fell to 6.45% in April, within the MPC majority’s 6-6.5% range for the medium-term equilibrium rate. The MPC believes that there is additional labour market slack associated with part-timers working fewer hours than they would like, although the implications of this shortfall for wage pressures is uncertain. With a further unemployment decline assured, the majority position suggest an interest rate rise before end-2014.
Governor Carney, of course, is strongly resistant to such a scenario, and may press for a downward revision to the estimated range for the medium-term equilibrium unemployment rate in the August Inflation Report, probably to 5.5-6%. This change would be presented as a response to new information; it would, in fact, simply make explicit an assumption already incorporated in the Bank’s May projections.