“Other assets” on the Bank of England’s balance sheet rose by a further £3.0 billion in the week to Wednesday 17th October, bringing the total increase since the run on Northern Rock started to £16.0 billion. This is the best available estimate of the extent of the Bank’s support to the lender.

Remarkably, the previous week’s estimate of £12.9 billion was confirmed by Northern Rock chairman Matt Ridley before the Treasury committee on Tuesday. Dr. Ridley stated that “the sums involved that have been reported of around £13 billion are approximately correct”.

The weekly increase of £3.0 billion is up from £2.3 billion and £2.9 billion in the prior two weeks, indicating that Northern Rock’s funding problems remain acute. There had been hopes that the rate of increase would slow following the Government’s announcement last week that its guarantee would be extended to new retail deposits.

As argued earlier, the Bank’s injection of funds into Northern Rock contributed to an easing of money market conditions, reflected in a fall in three-month interbank interest rates from a high of 6.90% on 11th September to a low of 6.24% on 5th October. However, the Bank has acted to reverse this effect in recent weeks by reducing the volume of lending to banks in its normal market operations and taking other steps to drain liquidity. The banking system’s reserves at the Bank have fallen from a peak of £30.0 billion on 25th September to £21.7 billion yesterday, while three-month interbank rates have firmed to 6.29% and remain unusually high relative to the 5.75% Bank rate. The Bank should arguably have allowed the higher level of reserves resulting from its support to Northern Rock to remain within the system until the interbank market had normalised.

The MPC voted 8-1 for unchanged rates in October (one vote for a 25 basis point cut) against our MPC-ometer’s 7-2 forecast. The one vote miss is in line with the model’s average forecast error over the last year. The minutes indicate that the Committee will give serious consideration to cutting rates in November. Three factors are likely to be important in determining the outcome.

The first is the preliminary estimate of GDP growth in the third quarter, released tomorrow. Empirical analysis shows that the early GDP estimate is a significant influence on MPC decisions. (This partly explains the tendency to move rates in the middle month of a quarter.) In its 10 year history, the Committee has never cut rates when quarterly growth has been 0.7% or higher. Indeed, there is only one example of a cut following a 0.6% GDP increase (in November 2001, just after the 911 terrorist attack).

The second key influence will be forward-looking growth components of business and consumer surveys for October released over the coming fortnight. The MPC is likely to give greater-than-normal weight to survey evidence given its belief that the “credit crisis” represents a negative shock to economic activity. September results were surprisingly resilient – most measures of confidence and future activity remain above their long-term averages. Significant weakening is necessary to justify a pre-emptive rate cut.

Thirdly, most MPC members will be keen to see greater evidence that inflation expectations have peaked before voting to cut rates. Business survey price balances were generally still strong in September and recent increases in wholesale oil and gas prices are concerning. As noted in the October minutes, the sharp fall in headline CPI numbers since the spring has yet to lead to any moderation in consumer inflation expectations.

I think the hurdles for a November cut are high but will continue to be guided by the MPC-ometer – it currently suggests stable rates but fewer than half of the components are available.

Bears of commercial property point out that current rental yields are low by historical standards. The CBRE all property prime yield edged up to 5.1% in the third quarter but remains well below its average of 6.4% over 1972-2006. However, the long-run average may be a poor guide to current “fair value”, for two reasons.

First, rents fluctuate significantly with the economic cycle. A high yield may not indicate that property is cheap if rents have been boosted above a sustainable level by a buoyant economy. Conversely, it may be right to invest when yields are low if rents are below trend and likely to benefit from future strong economic growth.

Secondly, any judgement about valuation must take account of returns on competing assets. The rental yield is often compared with yields on conventional gilts but this is invalid because bond interest is fixed while rents rise with inflation over the long run. In other words, the rental yield should be compared with real not nominal interest rates.

The chart shows a measure of valuation that incorporates these considerations – the gap between the normalised or cyclically-adjusted rental yield and real yields on long-term index-linked gilts. The normalised yield is currently lower than the actual yield (4.8% versus 5.1%) because rents are estimated to be 7% above trend, reflecting the economy’s recent strength. (There were much greater deviations in the early 1970s and late 1980s, when rents overshot by 30-40%.) However, the low level of the normalised yield is counterbalanced by similarly modest yields on index-linked gilts. The gap between the two is therefore only 20 basis points below its long-term average. In other words, based on current index-linked yields “fair value” for the actual rental yield is about 5.3% compared with the third quarter level of 5.1%.

Commercial property valuations are clearly much less compelling than in the 1990s but look defensible relative to bonds.

At the risk of committing heresy, I think the US housing market may have hit bottom. Needless to say, any revival would be taken badly by the Treasury market, which is counting on further Fed rate cuts to stave off a housing-led slump.

Home sales plunged over the summer as mortgage credit conditions tightened and the sub-prime crisis hit confidence. However, the home-buying conditions index of the University of Michigan consumer survey recovered sharply in early October, reaching its highest level since May. As the chart shows, this index tends to lead home sales.

Rather than extrapolate falling prices, consumers appear to be viewing recent weakness as a reason to buy. The share reporting favourable conditions because of low prices rose to a record high in early October. Stable labour market conditions and falls in mortgage rates for prime borrowers may also have contributed to the improvement in sentiment.

Any recovery in demand could quickly boost homebuilding activity. Single-family housing starts have fallen by much more than new sales since the market peaked in late 2005. New construction has been below the pace of sales since July last year and the number of unsold homes declined to a 19-month low in August.

I could be early. This week's release of the NAHB homebuilders survey for October and housing starts and permits for September will provide further information. Homebuilders' stock prices are still languishing near their lows, so have yet to signal an improving outlook.