Category: Money Moves Markets

In December last year the Council of Mortgage Lenders (CML) predicted that the percentage of mortgages more than three months in arrears would rise from 1.8% at the end of 2008 to 4.4% by end-2009. First-half performance has been much better than expected, with the arrears proportion standing at 2.4% at the end of June.

As well as undershooting forecasts, current arrears experience compares favourably with the recession and housing market downturn of the early 1990s. The CML’s series for three-month-plus arrears starts in 1995 but rough estimates for earlier years can be derived from data on six-month-plus cases. The current 2.4% arrears proportion compares with an estimated peak of about 6% in 1992 – see chart.

As pointed out on page 29 of the August Inflation Report, employment has fallen by less than at the comparable stage of the last downturn, partly reflecting cuts in real pay. Government schemes are also helping: 220,000 households were receiving income support mortgage interest payments in May (although such support was also available in the 1990s), while the Department of Communities and Local Government has estimated eventual take-up of the homeowners mortgage support scheme – which allows borrowers to defer interest payments for up to two years – at 42,000.

The key factor suppressing arrears, however, is a lower burden of interest service costs than in the early 1990s. Household interest payments peaked at 10.9% of disposable incomes in the fourth quarter of 2007 versus a high of 15.0% reached in the third quarter of 1990. One year into the 1990-91 recession, the interest burden was still 11.6% of income; the latest figure – for the first quarter of 2009 – is 8.7% and a further decline to 7-7.5% is likely, based on more recent Bank of England data on effective interest rates.

So the interest burden is now nearing the bottom of its historical range, despite a record level of debt. A bearish view is that mortgage defaults have simply been postponed because interest service will rise rapidly once official rates start to normalise. The impact of policy tightening, however, may be partly offset by a narrowing of current wide lending spreads. Moreover, MPC action will be conditional on a solid recovery, implying more favourable labour market conditions for borrowers.

(Please note: an earlier version of this post used a National Statistics series for household interest payments that nets off an estimate of consumption of financial intermediation services. The chart and text have been updated to include these payments in the analysis, as is appropriate. Thanks to an observant reader for spotting this omission.)

Professor Charles Goodhart has advocated charging banks for holding reserves at the Bank of England to encourage them to lend out the cash created by quantitative easing. He cites the example of Sweden, where the interest rate on the Riksbank’s deposit facility has been set at minus 0.25%.

Current Swedish monetary policy and money market arrangements, however, are not comparable with the UK’s. The Riksbank has cut its target interest rate, the repo rate, to 0.25% but has not engaged in quantitative easing, in the sense of asset purchases financed by creating new reserves. The deposit rate has fallen to -0.25% because it has recently been set 50 basis points below the repo rate.

Swedish banks make little use of the deposit facility because they are able to lend to the Riksbank in daily “fine-tuning” operations at the repo rate minus 10 basis points – i.e. 0.15% at present. The Riksbank’s weekly statement shows that these fine-tuning loans currently stand at SEK176 billion versus deposits of only SEK39 million. So the negative deposit rate has little practical relevance.

In the UK, a key objective of market operations is to maintain overnight interest rates in line with Bank rate. If the Bank of England stopped paying interest on reserves, banks would attempt to earn a return on their cash by lending it out short term in secured money markets; this increased supply would push overnight rates close to zero. In other words, the change would undermine the anchor role of Bank rate and, by extension, the MPC’s control over monetary conditions.

On the same logic, if the Bank charged banks for holding reserves, overnight rates would turn negative. Banks would then have an incentive to hold liquidity in the form of bank notes, which would at least maintain a fixed value, rather than deposits at the central bank or overnight loans. To prevent such behaviour, the Bank would have to place restrictions on banks’ ability to convert their reserve holdings into bank notes – another fundamental feature of current monetary arrangements.

Paradoxically, banks as a group would be unable to avoid charges even if they increased their lending as desired, since the aggregate amount of reserves would be unaffected. While an individual bank might succeed in reducing its deposits, other institutions would find themselves holding more cash. The aggregate level of reserves is effectively fixed by the Bank of England, assuming a stable demand for bank notes.

Recent lending stagnation has been partly due to a fall in credit demand. To the extent that supply of loans is constrained, this reflects banks’ efforts to conserve capital and reduce risk, rather than the competing attraction of earning 0.5% by holding cash at the Bank of England. Even if the technical obstacles could be overcome, a reserves-charging scheme would probably have little impact on lending behaviour.

As expected, food prices had a favourable impact on July consumer prices, cutting the annual increase by 0.14%. This effect, however, was offset by a pick-up in “core” inflation, resulting in the headline CPI rise remaining at 1.8%. This is well above the Inflation Report forecast of average inflation of about 1.25% in the third quarter and casts doubt on Bank of England Governor Mervyn King’s suggestion of a fall below 1% later in 2009.

The CPI excluding food, energy, alcohol and tobacco rose by an annual 1.8% in July, up from 1.6% in June and the highest since November. Based on recent National Statistics research, this measure of core inflation would probably stand at 2.4-2.5% in the absence of December’s VAT cut.

Stubborn core trends partly reflect the continuing impact of last year’s sterling depreciation but also call into question consensus and Bank of England estimates of the sensitivity of inflation to rising economic slack. An alternative forecasting approach based on monetary growth may be more consistent with recent numbers than “output gapology” – see previous post.

Retail prices were down by an annual 1.4% in July but this compares with a 1.6% June decline. Interestingly, the housing depreciation component of the RPI – linked to house prices – rose last month for the first time since June last year. As argued in the previous post, RPI inflation should rebound strongly and exceed CPI inflation in 2010 as housing and interest rate effects unwind.

Energy price falls in late 2008 and early 2009 helped to offset the impact of a decline in labour incomes on consumer spending power. Energy prices have started to firm again recently but household budgets have enjoyed relief from a new source – a decline in food bills.

A surge in consumer food prices during 2008 squeezed real incomes and contributed to a cut-back in household spending. Food commodity prices, however, weakened sharply late last year and this decline has been feeding through at the retail level recently, with food components of consumer price indices in the US, Japan and Euroland recording an unusual fall over the last six months – see chart.

Food commodity prices recovered during the first half of 2009 but the Economist food price index is currently still 26% below the peak reached in July last year in dollar terms. Despite a headline-grabbing spike in sugar prices, the index has retreated 5% from a recent high in early June. So food CPI trends may remain subdued during the second half.

UK food inflation peaked at a higher level and has been slower to subside, reflecting sterling’s big decline during 2008. With the pound recovering recently, however, this effect is reversing and food prices should suppress CPI numbers over coming months (producer prices suggest a favourable impact in the July report released tomorrow – see chart in previous post).

Major British banks (i.e. the high-street groups covered by British Bankers’ Association statistics) incurred impairment charges of £31 billion in the first half of 2009, up from £26 billion in the second half of 2008, according to recent interim statements. As explained below, a comparison with the bad debt cycles of the early 1980s and early 1990s suggests further provisions of £100-150 billion over the three and a half years to the end of 2012.

More optimistically, the same group of banks made pre-impairment operating profits of £26 billion during the first half. Providing this run-rate is maintained, cumulative profits between the second half of 2009 and 2012 should be sufficient both to cover future impairments and allow some addition to capital, implying no need for a further government “bail-out”.

Loss provisions by major banks totalled 9% of credit exposure over the five years from 1982 to 1986, following the 1979-81 recession, and 7% over 1989-93, encompassing the 1990-91 recession. The 1982-86 figure includes losses on sovereign loans, although these were not formally recognised until later in the 1980s. (See previous post for more details.)

A similar loss rate, i.e. between 7% and 9% over five years, is plausible in the current cycle. It is not clear that recent bank behaviour was any more reckless than in the rush to lend to developing countries in the late 1970s or the property lending boom of the late 1980s. Moreover, nominal and real interest rates are much lower than in the 1980s and 1990s, which should limit defaults.

Based on 2008 exposure, a 7-9% rate of attrition would imply credit losses of £190-240 billion over the five years 2008-12. Impairment charges totalled £64 billion in 2008 and the first half of 2009 so this suggests losses of £125-175 billion over the three and a half years to the end of 2012.

In addition to impairment charges, however, Lloyds, HBOS and RBS recorded a combined net trading loss of £23 billion in the 18 months to mid 2009, reflecting writedowns of securities. On the basis that banks held a much greater proportion of credit exposure in securitised form in the current cycle, such writedowns should be included in total losses recognised to date when comparing with the early 1980s and early 1990s.

A conservative approach is to incorporate only the net trading loss of these three banks, rather than aggregate trading writedowns. (Writedowns by other banks have been offset by income from other trading activities.) Assuming no further trading losses, this reduces the forecast range for provisions to £100-150 billion between the second half of 2009 and 2012.

The chart translates this range into a half-yearly profile, assuming that impairments decline steadily from a peak in the first half of 2009. The middle of the range implies annual totals of £57 billion, £40 billion, £23 billion and £6 billion over 2009-12.

Banks’ pre-impairment profits of £26 billion during the first half of 2009 incorporate £6 billion of trading gains by Barclays and RBS, offset by a further £3 billion loss by Lloyds / HBOS. Investment bank profits could be less favourable going forward but any decline should be offset by higher net interest income as average lending / deposit rate spreads recover from recent lows – see previous post on banks’ margins.

If pre-impairment profits are stable at £26 billion per half-year – arguably cautious – and impairment charges total £100-150 billion between the second half of 2009 and 2012, banks will earn cumulative post-provision profits of £30-80 billion by the end of 2012. This surplus will be available to boost capital levels, on top of any addition from further fund-raising in markets.

The August Inflation Report is much less dovish than the market expected and signals that the Monetary Policy Committee now has a neutral policy bias, following the £50 billion expansion of QE announced last week.

The mean inflation forecast based on an unchanged level of Bank rate is an estimated 2.1% and rising at the two-year horizon. This is up from 1.7% in May and a record low 0.4% in February, and the first above-target projection since August last year. The deviation from 2% is an indicator of policy bias and has often signalled rate moves – see chart.

This inflation view rests on a respectable economic recovery but the MPC’s GDP forecasts are not particularly aggressive. Output rises by about 2% in the year to the second quarter of 2010 in the central case based on unchanged rates. Risks, however, are weighted to the downside, so the mean projection is only about 1.5%.

In addition to the two-year-ahead projection, the MPC has raised its shorter-term inflation forecasts, partly reflecting a less optimistic assumption about domestic energy prices (expected to fall by 5% during the second half versus 15% in the May Report). After an undershoot in the second half, inflation returns to about 2% in early 2010.

Bank of England Governor Mervyn King referred to the likelihood of having to write a letter to the Chancellor explaining a fall in inflation to below 1% later this year. However, the central-case and mean projections for the third and fourth quarters are above 1%, suggesting that any move below will last a single month (and would not have occurred without the VAT cut).

Echoing earlier remarks in a speech, Deputy Governor Bean referred to the option of achieving a future tightening of policy initially by raising Bank rate, with gilt sales staggered over a longer period and conditioned on market developments. The difficulties of reversing QE imply that rate rises, when they begin, could be rapid.

Unemployment is still rising rapidly but job vacancies – a coincident rather than lagging indicator – suggest economic stabilisation. Vacancies bottomed in May and have edged higher in June and July. The rate of change over three months is still negative but has recovered to a level historically consistent with marginal GDP expansion – see chart.

The May trough in vacancies, like the peak in February 2008, was signalled in advance by the Market job placements index – see earlier post. The Markit index fell back in July but remains well above the low reached in December 2008.

US labour market statistics for July suggest that the recession has ended, for two reasons. First, the (smaller) fall in payroll employment last month was offset by a rise in the length of the average working week. Aggregate hours worked in the private sector were therefore unchanged – the first month not to register a fall since last August. Since labour productivity has continued to rise during the recession, stability in hours worked implies an increase in output.

Secondly, the unemployment rate – derived from a survey of households rather than employer payrolls – fell slightly in July. Historically, a monthly decline following a sustained steep increase has marked the end of a recession – see chart. This is true even when the rate subsequently rises to a new peak, as it did in the aftermath of the 1990-91 and 2001 recessions.

Some commentators have dismissed the significance of the fall in the unemployment rate on the basis that it reflected a contraction of the labour force rather than a rise in the number of households in employment. This may be unwise. Historically, the first monthly decline after a significant peak has often been associated with a fall in the labour force, e.g. in 1975, 1980, 1982 and 2003. Moreover, a measure of payroll jobs derived from the household survey increased, by 70,000, last month.

The improvements in both the employer and household surveys are consistent with other labour market evidence, including a recent large decline in corporate layoffs (see here), falling initial unemployment claims, a stabilisation of help-wanted advertising and less negative employment readings in business surveys.

OECD trade – the combined volume of exports and imports – contracted by 18% between the second quarter of 2008 and the first quarter of this year. However, two pieces of information released this week suggest that trade is now recovering.

First, German export orders rose by a further 8% in June, to stand 19% above the low reached in February. Orders are correlated with OECD trade volumes, with a historical “beta” or elasticity of about two – see first chart.

Secondly, the sum of the US ISM manufacturing imports and export orders indices broke above 100 in July, i.e. more firms now report rising volumes than falls. This indicator is also a good proxy for changes in OECD trade – second chart. The rise in the imports index probably mainly reflects slower destocking by US firms.

A comparison with the last three recessions suggests that National Statistics’ current estimate of a fall in GDP of 5.7% between the first quarter of 2008 and the second quarter of 2009 will be revised to show a significantly smaller decline over coming years. This claim is supported by labour market indicators, which, though weak, have deteriorated by less than would have been expected given the estimated GDP drop.

The Bank of England has helpfully compiled a “real-time” database of national accounts statistics that, for some series, including GDP, extends back to 1976. Following the end of the recessions in 1974-75, 1979-81 and 1990-91, GDP was estimated to have declined from peak to trough by 4.6%, 6.5% and 4.3% respectively. In the latest vintage of statistics, the falls are 2.7%, 6.0% and 2.5%. So revisions have cut the GDP drop by 1.9 percentage points for 1974-75, 0.5 pp for 1979-81 and 1.7 pp (after rounding) for 1990-91.

The smaller adjustment in 1979-81 may be misleading because revisions have also resulted in a major change to the profile of the recession. The originally-estimated 6.5% GDP decline referred to the change between the second quarter of 1979 and the third quarter of 1981 but the recession trough was subsequently shifted to the first quarter of the latter year. The latest statistics show a 4.6% GDP decline over the original recession period.

These comparisons indicate that the current estimate of a 5.7% GDP decline by the second quarter of 2009 could eventually be revised down by as much as 1.9 percentage points. A simple model for GDP growth based on changes in vacancies and claimant-count unemployment confirms that a substantial adjustment is possible. The model tracks the GDP falls in the last three recessions reasonably closely and suggests an annual decline of 3.7% in the first quarter of 2009 versus a current official estimate of 4.9% – see chart.