Few economists are talking about a possible near-term inflation rise – see Tuesday's post – but it is on the Bank of England's radar, judging from a speech this week by MPC hawk Andrew Sentance:

We have seen two major spikes in CPI inflation within the last two years. And it is possible that another one is in prospect when higher VAT comes into effect early next year, particularly if this is accompanied by a continuation of the recent surge in energy and commodity prices. These repeated episodes risk eroding confidence that UK policy-makers remain committed to low and stable inflation, with a knock-on effect on inflation expectations in financial markets, among the business community and the public which could be self-fulfilling.

Equity market valuation is often assessed using Professor Robert Shiller's "cyclically-adjusted price-earnings ratio" (CAPE), which divides the real (i.e. inflation-adjusted) index level by a 10-year one-sided moving average of real earnings per share. The smoothing is an attempt to estimate trend earnings, with 10 years consistent with the typical 7-11 year length of the dominant economic cycle (the Juglar business investment cycle).

For the S&P 500, Shiller's CAPE stood at 21.0 in early October versus a median since 1881 of 15.7, a 34% premium, suggesting that equities are significantly overvalued relative to history – see first chart.

Because the moving average is one-sided, however, it can under- or overstate trend earnings when the growth rate is changing. S&P 500 trend earnings expansion may have accelerated in recent years, partly reflecting a stronger contribution from foreign profits related to US companies' exposure to emerging economies.

A further problem at present is that, unusually, the 10-year moving average incorporates two major earnings recessions, reflecting the proximity of the 2001 and 2008 economic downturns. These two factors suggest that Shiller's CAPE is overstating equity market valuation.

The first issue can be addressed by using a two-sided moving average, so that the trend measure incorporates future as well as past data. The average, of course, cannot be calculated for recent years (the last five years in the case of a 10-year window). Trend, however, can be estimated using consensus earnings forecasts and extrapolation – the estimates will be subject to revision but the procedure may produce superior results to one-sided smoothing.

This two-sided approach also solves the current problem of the moving average incorporating two earnings recessions, since the 2001 downturn is no longer within the time window of the calculation.

The second chart shows S&P real earnings per share, expressed at 2010 prices, together with an estimate of trend calculated on this basis. Earnings are back at trend following an unprecedented collapse in 2008-09.

The third chart shows a price-earnings ratio calculated using the trend earnings measure. The current level of 16.3 compares with a median since 1881 of 14.5 – still 12% higher but a much smaller deviation than for Shiller's CAPE.

Comparing current valuations with a median extending back to the late 1800s is questionable – the equity risk premium may have fallen, reflecting such factors as greater diversification possibilities, improved liquidity and lower trading costs. As the second chart shows, a trend line through the data has a mild upward slope. If this line, rather than the median, is used to gauge sustainability, equities are 8% undervalued, based on a current price-trend earnings ratio of 17.7.

The economic scenario viewed here as most likely involves the current "soft patch" in global industrial activity continuing into the autumn but falling short of a "double dip", with a revival of momentum following in early 2011. Recent news seems consistent with this script.

The first chart below shows six-month growth rates of G7 plus emerging E7 industrial output and a composite leading index derived mainly from OECD country indices. A "leading indicator of the leading index" is also plotted, based on its shorter-term momentum.

The leading index itself continued to slow sharply in August and its recent weakness is now being reflected in industrial output. The double-leading indicator, however, moved up from zero, suggesting that six-month growth in the index will bottom at this level within the next couple of months, with output itself following around end-2010. The rise was marginal but the indicator is a smooth series and usually gives reliable signals of turning points.

An upturn from late 2010 would be consistent with recent better G7 monetary trends: six-month growth in real M1 troughed in January 2010 and led the leading index and industrial output by 8-9 months and 11 months respectively at the last two turning points – see previous post. Available September figures suggest that the real M1 pick-up has been sustained, although the rate of expansion is much lower than in late 2008 and early 2009 – second chart.

What could go wrong? QE2, arguably, poses the greatest risk – any direct economic boost from a further monetary injection is likely to be outweighed by associated strength in commodity prices, which will cut real incomes and increase pressure for E7 policy tightening. Optimists should hope that the Fed reins back on its plans – markets would suffer a short-term setback but the price would be worth paying for a more stable economic outlook.

Regular pay rose by an annual 2.3% in August, the fastest for 15 months. A measure of total pay growth, incorporating a 12-month moving average of bonuses, was 2.2%, a 19-month high – see first chart.

The annual rates of change of regular and total pay bottomed in late 2009 a few months after the retail prices index (RPI), which had been pushed down by interest rate cuts, house price falls and the temporary reduction in VAT. Current pay growth remains unusually low relative to RPI inflation, suggesting a further increase. Bonuses, in particular, could strengthen further, given buoyant corporate profits.

Elsewhere in today's labour market report, job vacancies fell further in September, suggesting that the recent recovery in employment will stall. A more positive message, however, emerges from the latest Monster survey of on-line recruitment activity: the employment index, which correlates with and may lead the official vacancies series, rose further last month – second chart. (The survey measures opportunities posted on corporate career sites and job boards, including Monster. The index levels in the chart have been seasonally adjusted.)

QE2 is damaging global economic prospects as an associated rise in commodity prices squeezes real incomes and forces central banks in emerging economies to tighten policies.

China's decision this week to raise reserve requirements temporarily for six large commercial banks suggests that another phase of restriction is beginning in response to economic reacceleration, rising industrial cost pressures and soaring food prices – see previous post.

Based on recent gains in the weekly food produce price index, the CPI food component could rise by about 3% between August and October – see chart. Since it was little changed between August and October 2009, this would push the annual increase up from 7.5% to more than 10%. With food accounting for one-third of the CPI basket, this, in turn, implies that headline inflation could rise from 3.5% in August towards 4.5%.

CPI inflation was stable at 3.1% in September but may rise towards 4% over coming months, reflecting surging commodity prices and the VAT hike. This would squeeze consumer budgets, thereby increasing the risk of a "double dip", while making it difficult for the MPC to launch QE2. There is a strong case for postponing the VAT increase until external inflationary pressures ease.

Key factors suggesting a rise in inflation include:

• VAT pass-through. The planned hike in the standard rate from 17.5% to 20% in January 2011 follows a rise from 15% this January so might be expected to have little impact on annual inflation. The Bank of England's regional agents' survey, however, suggests that most firms plan to pass on the increase in full, in contrast to the 2010 rise (which was a reversal of a temporary reduction). Assuming that pass-through was 50% this year and will be 90% for the coming hike, the annual CPI rate would be boosted by 0.6 percentage points.

• Food prices. A post a month ago suggested that CPI food inflation would rise to an annual 7% but a further substantial increase in commodity prices in recent weeks could result in a move to 10% (below the 14.5% peak reached in 2008). From 4.9% in September, this would add 0.5 percentage points to headline inflation, given food's 9.6% weight. Additional upward effects are possible via the beverages and catering services CPI components.

• Household gas bills. Ofgem recently estimated that energy suppliers' gas purchase costs would increase by 13% by next spring, based on higher forward prices. The impact on household gas bills is unclear but a 10% rise, in contrast to a 5.7% cut over the last year, would add 0.4 percentage points to annual CPI inflation, given a 2.5% weight.

• Motor fuel prices. Fuel cost inflation has fallen significantly since the spring and should decline further as price hikes in late 2009 and early 2010 drop out of the annual comparison. Pressures are reemerging, however, with a recent rise in wholesale costs, on top of this month's duty increase, suggesting that a litre of unleaded will return to about £1.20, from £1.15 in September. Fuel inflation, therefore, will subtract less than previously expected from the headline CPI rate – an estimated 0.3 percentage points by early 2011.

In combination, these factors imply a boost of 1 percentage point or more to annual CPI inflation by early next year. A rise, therefore, looks inevitable, even assuming some slowdown in other CPI components.

One further point: CPI inflation continues to be suppressed by a suspiciously-low estimate of clothing and footwear price rises. While the RPI clothing and footwear index rose by an annual 9.4% in September, the corresponding CPI increase was just 0.9% – National Statistics, in effect, assumes that consumers are so expert at shopping around that they are able to buy the same volume for just 0.9% more than a year ago despite a 9.4% increase in label prices.