Blog

Industrial output in the Group of Seven (G7) major economies fell an estimated 12% between February and December last year, based on data for six of the seven countries (Canada has yet to publish for December). This is equal to the peak-to-trough decline during the 1974-75 recession – see first chart.

Business surveys signal a further fall in output in early 2009, implying the current G7 industrial slump will soon be the deepest since World War 2.

For a longer perspective, the second chart shows estimated annual average growth rates of G7 industrial output back to the late nineteenth century. The underlying country statistics were assembled from various sources, including an Economist publication One Hundred Years of Economic Statistics.

If annual average G7 output in 2009 were to equal the December level, it would be 9% lower than in 2008 (red line in chart). This would be the largest annual decline since 1938, when the “Roosevelt recession” in the US led to a 10% drop.

However, the current downturn would have to extend hugely in magnitude and duration to be comparable with the depression of the early 1930s. G7 industrial output plunged for three successive years – 9% in 1930, 13% in 1931 and 10% in 1932.

The third chart shows annual rates of change of G7 output and inflation-adjusted broad and narrow money supply measures. While G7 activity will slide further in early 2009, strengthening real money growth continues to offer hope of an improvement in economic trends later in the year.

Some analysts are wrongly claiming that the Bank of England’s asset purchase facility will have no monetary impact. For example, an article in Friday’s Financial Times stated that “the money used to buy the corporate securities will be financed by Treasury sales of government bills rather than the creation of money”.

The simultaneous sale of Treasury bills will ensure no effect on banks’ reserves held at the Bank of England and hence the monetary base M0, implying the scheme does not represent “quantitative easing” in the Japanese sense. Providing Treasury bills are sold to banks, however, while the Bank purchases assets from non-banks, the broad money supply will expand. Banks should indeed absorb much of the increase in the supply of bills, given regulatory pressure to raise their holdings of safe liquid assets.

The potential for the scheme to boost M4 directly is more important than whether it expands the monetary base. M0 growth has already risen significantly as a result of increased Bank of England lending to the banking system, without any noticeable impact on broad money or credit conditions. In the US, wider money measures have accelerated since the Federal Reserve began buying commercial paper and agency securities, having shown little response to earlier initiatives inflating the monetary base.

Another article in the same FT edition claims that quantitative easing can begin only after interest rates have been cut to zero; otherwise “the overnight interest rate in money markets would in any case fall towards zero because commercial banks would find themselves awash with unwanted cash looking for a home overnight”. This is incorrect. Banks are able to deposit excess funds in unlimited amounts at the Bank of England’s operational standing deposit facility, earning interest at Bank rate minus 25 basis points, a level that sets an effective floor for overnight rates.

With the option of underfunding the budget deficit apparently still off the policy agenda, the asset purchase facility represents the best hope for an early and significant revival in broad money growth. Rapidly implemented and suitably expanded, the scheme could lay the foundations for an economic recovery from late 2009.

The tier 1 capital ratios of major British banks would fall below the 6-7% minimum required by the Financial Services Authority if they were forced to write down the value of their mortgage portfolios in line with the market prices of mortgage-backed securities used by the Bank of England in its operation of the special liquidity scheme (SLS). However, actual losses are likely to be a fraction of those implied by these prices, illustrating the absurdity of mark-to-market assessments of capital adequacy.

When the SLS closed on 30 January, the Bank of England held securities with a nominal value of £287 billion as collateral against Treasury bills lent under the scheme. The Bank’s valuation of these securities was £242 billion, implying a discount to par of about 16%. The collateral was mostly in the form of AAA-rated residential mortgage-backed securities and covered bonds. Since AAA tranches suffer impairment only after lower-rated tranches have been wiped out, a 16% discount suggests a much larger expected loss – of perhaps 25% – on the underlying pool of mortgages.

Major banks held £496 billion of residential mortgages on their balance sheets at the end of 2008, according to the British Bankers’ Association. Ignoring additional exposure via off-balance-sheet entities, a write-down of 25% would reduce capital by about £125 billion – sufficient to cut banks’ current aggregate tier 1 ratio of over 11% by more than half.

The chances of actual losses on this scale are miniscule. In the worst year of the early 1990s housing downturn – 1991 – 0.77% of mortgaged properties were repossessed, according to the Council of Mortgage Lenders. Even assuming a repossession rate of 0.77% sustained for 25 years, and a loss given default of 50%, the cumulative reduction in the value of mortgage principal would be less than 10%.

The large deviation of market prices of mortgage-related securities from their likely economic value reflects extreme investor risk aversion and illiquidity. There is a strong case for the Bank of England to use its new asset purchase facility to buy such securities, in addition to corporate bonds and commercial paper. Targeting a wide range of assets would facilitate an early expansion of the facility – necessary if it is to have a meaningful impact on monetary growth.

Sharp falls in annual consumer and retail price inflation in December have prompted claims that the economy stands on the brink of deflation, justifying today’s further reduction in Bank rate to 1.0%. The declines, however, are entirely explained by the 2.5 percentage point cut in VAT, falling energy costs and – in the case of the RPI – lower interest rates. Underlying inflation has yet to show much response to economic weakness.

Annual CPI inflation fell from 4.1% in November to 3.1% in December. Since its first statistical release a fortnight ago, the Office for National Statistics (ONS) has published a December number for the “CPI at constant tax rates” (CPI-CT), i.e. adjusted for the impact of the VAT cut. This incorporates official estimates of the extent to which the reduction was passed on by retailers and other consumer suppliers. Annual CPI-CT inflation actually rose in December, from 3.9% to 4.1% – see chart.

The annual increase in the CPI’s energy component moved lower in December, from 16.7% to 12.2%, mainly reflecting falling petrol prices. If the rate of change had remained at 16.7%, annual CPI-CT inflation would have risen further, to an estimated 4.3%.

To gauge underlying pressures, it is helpful to exclude both energy prices and the effect of the VAT reduction. The annual increase in the CPI excluding energy eased from 3.1% to 2.3% in December. Without the VAT cut, however, the December number would have risen to an estimated 3.4% – equal to the peak reached in August / September.

Annual RPI inflation dropped from 3.0% in November to just 0.9% in December. The ONS does not publish an RPI series at constant tax rates but RPIY – which excludes indirect taxes and mortgage interest costs – rose an annual 3.9% in December, unchanged from November.

The severe recession will rapidly reduce domestic inflationary pressures but the slump in sterling is having an offsetting impact, with manufactured import prices rising by an annual 14% in November (December figures are due next week). Energy price effects will ensure a further significant drop in headline CPI and RPI rates during 2009 but the decline is likely to be smaller than the MPC and consensus expect, while VAT-adjusted non-energy inflation may remain above 2%.

My MPC-ometer projects a further half-point cut in Bank rate to 1.0% tomorrow. This is also the expectation of 61 out of 68 economists, according to a Reuters survey. The model’s forecast is heavily influenced by the 1.5% slump in GDP in the fourth quarter reported a fortnight ago. Further falls in consumer confidence and manufacturing pricing plans and January’s weak stock market performance are also contributory factors.

In my view, another cut in rates is less important than measures to boost underlying broad money growth, which remains too low to support an economic recovery – see here. Improving credit availability is clearly also a priority but credit constraints would be relieved by an increase in the quantity of money circulating in the economy.

The simplest way to boost broad money would be to “underfund” the huge budget deficit, i.e. finance it by borrowing from banks or the Bank of England rather than by issuing gilts. This would not imply a reduction in banks’ capacity to lend to firms and households, because institutions are not required to set aside additional capital against increased lending to the government. Underfunding is a conventional unconventional measure, to use Bank of England Governor Mervyn King’s terminology. It occurred after the early 1990s recession, helping to underpin an economic recovery, and again in 2001-02, when the UK managed to skirt a global downturn.

An alternative to underfunding would be for the Bank of England to use its asset purchase facility to buy securities from non-bank financial institutions and companies. Such purchases would directly boost broad money; by contrast, buying assets from banks has a positive impact only if they respond by increasing lending. The Bank could, for example, buy newly-issued commercial paper or longer-maturity asset-backed securities (more likely to be held outside the banking system).

Whether such transactions boost the monetary base (i.e. currency in circulation and banks’ deposits at the central bank) – thereby qualifying as “quantitative easing” – is of secondary importance. Banks’ unwillingness to lend reflects capital constraints and risk aversion rather than a shortage of cash. Monetary base growth has already picked up significantly as a result of the Bank of England’s expanded lending to the banking system, without any noticeable impact on the broad money supply or credit availability – see chart.

A further interest rate cut could be counterproductive in terms of improving credit availability. Banks need to be able to widen interest margins in order to rebuild capital to support higher lending. As the general level of interest rates approaches zero, however, their ability to lower deposit rates is constrained by increasing competition from government-guaranteed savings products offering full security with little or no loss of income. Forced to lower lending rates to existing borrowers on deals linked to Bank rate or LIBOR, banks may compensate by raising rates charged on new loans.

The Bank of England yesterday reported that banks and building societies had borrowed £185 billion of Treasury bills under the special liquidity scheme (SLS) by the time of its closure on 30 January – broadly consistent with my earlier estimates (e.g. here). The SLS has been superseded by the Bank’s new discount window facility (DWF). Drawings from the DWF will be published quarterly but with a considerable lag – first-quarter figures will be available at the end of June.

The net percentages of senior bank loan officers reporting tighter credit standards on loans to businesses and residential mortgages declined between October and January, according to the Federal Reserve’s latest survey – see first chart. However, the balances remain above the peaks reached in the 1990-91 and 2001 recessions.

The second chart shows the annual rate of change of industrial output and an average of the net percentages of loan officers reporting tighter standards on loans to large / medium and small firms, plotted inverted. Turning points in the latter series lead the industrial cycle. The latest small change is consistent with an approaching trough in the annual rate of decline of output, confirming the message from real money trends – see here.

British-owned banks are being unfairly blamed for a contraction of sterling banking activity by foreign-owned institutions operating in the UK.

Total sterling assets of UK-based banks grew by 8% in the year to December, down from 14% in the prior 12 months, according to Bank of England data. The 8% rise, however, conceals a 15% increase for domestically-owned banks offset by a 3% contraction in sterling assets of foreign institutions, which account for 34% of the amount outstanding.

Foreign banks continued to expand their lending to the non-bank private sector last year, though at a slower pace than British-owned banks. The decline in their total sterling assets reflected a cut-back in lending to other UK banks – this in turn constrained the ability of British banks to extend new credit to households and firms.

Of the 34% foreign share of total sterling assets, “other EU” banks account for 22 percentage points, “other developed countries” 7 pp and the US 4 pp. Last year’s contraction was due to a large fall in assets of banks from “other developed countries”, a category including Swiss, Canadian and Australian institutions, among others. Sterling assets of “other EU” banks grew by 7% during 2008, partly reflecting Santander’s acquisition – via Abbey – of Bradford & Bingley’s savings business. Icelandic banks are also included under “other EU” – their liquidation could affect future data.

Detailed monetary statistics for December contain some glimmers of hope, suggesting the recession will hit bottom by the third quarter of 2009. In contrast to the US, however, UK monetary trends are not yet consistent with an economic recovery, in the sense of a return to trend growth or higher.

Annual rates of change of inflation-adjusted broad and narrow money have recovered from their October low, reaching the highest level since May – see first chart. This reflects slowing retail prices rather than much change in nominal money trends but is nonetheless meaningful – real money typically leads the economy by about six months. A further plunge in RPI inflation should sustain the revival in early 2009 but real growth rates would need to rise to 5-10% to signal an economic recovery.

A further positive development was a 1.8% monthly rise in broad money holdings of private non-financial corporations – the first increase since July and the largest since April 2007. The corporate liquidity ratio – money holdings divided by bank borrowing – could be bottoming, in contrast to a continued slide in the Eurozone – second chart. However, it remains very low by historical standards and usually surges ahead of recoveries.

Rises in net mortgage lending (from £830 million in November to £1.9 billion in December) and the number of mortgage approvals (from 27,000 to 31,000) are small and should be treated with caution. The increase in lending is explained by a fall in repayments – gross advances continued to slide to a seven-year low – while approvals were still down 58% from a year before.

In other news today, Japan’s industrial output plunged by 9.6% in December and – even more shockingly – manufacturers plan to slash production by a further 13% in January and February, according to a survey by the Ministry of Economy, Trade and Industry. Based on December data for the US and Japan and November numbers for other countries, G7 industrial output is now an estimated 10-11% below its February 2008 peak. With further weakness highly likely, the cumulative fall should soon challenge the 12% peak-to-trough decline in the brutal 1974-75 global recession – third chart.

The IMF’s new forecasts show UK GDP declining by 2.8% in 2009 versus an average fall in advanced economies of 2.0%. The UK’s performance is projected to be the worst of the Group of Seven (G7) major economies.

This forecast, however, is more a reflection of recent developments than a view that the UK will underperform going forward. UK GDP and output figures published last week imply that monthly GDP in December was 1.0% below the fourth-quarter level and 2.4% lower than the calendar 2008 average – see here. So the IMF forecast for 2009 implies a further decline in GDP from December of only 0.4%.

The UK is the only G7 country yet to have published fourth-quarter GDP numbers. Performance in other major nations is likely to be similarly dire or even worse, in which case IMF projections may need to be revised down.

After the UK, Japan and Germany are forecast to suffer the largest GDP declines in 2009, of 2.6% and 2.5% respectively, reflecting their exposure to collapsing world trade. Countries that “fixed the roof” in good times – running savings surpluses and avoiding domestic credit / housing booms – are suffering at least as much as deficit offenders in the current global downturn.

The Bank of England has estimated that the five largest UK banks and Nationwide would sustain aggregate credit losses over five years of £115-130 billion in a “severe but plausible macroeconomic risk scenario” (see the October Financial Stability Report, pp.28-9). As explained below, an examination of credit impairment suffered by major high-street banks in the aftermath of the recessions of the early 1980s and early 1990s suggests a larger five-year loss, of £190-240 billion. Even this amount, however, could be absorbed by existing capital resources and future profits, implying additional government financial support may be unnecessary.

Credit losses suffered by British banks in the 1980s reflected both a severe domestic recession at the start of the decade and the Latin American debt crisis, initiated by Mexico’s default in 1982. Banks had stepped up their lending to developing-country governments in the late 1970s as they sought to recycle surplus oil funds – comparable perhaps with their recent disastrous foray into US subprime mortgages.

International losses were less important in the early 1990s but domestic default experience was worse than in the 1980s, despite a less serious recession, probably because of punishingly high real interest rates necessitated by sterling’s membership, until September 1992, of the exchange rate mechanism. By contrast, current low nominal and real interest rates will enable some default-prone borrowers to continue to service their loans.

On this basis, it is defensible to assume that current credit losses will be similar to but no greater than those suffered in the 1980s and 1990s downturns. One difference from these earlier episodes is that a much greater portion of banks’ credit exposure now resides in their securities portfolios rather than the traditional loan book. So it is appropriate to compare loan losses in the earlier recessions with total losses – both loan impairments and writedowns of securities – in the current cycle.

The chart shows high-street banks’ annual “impairment and other provisions (net)”, as reported by the British Bankers’ Association (BBA), expressed as percentage of their “assets at risk”, defined here as total assets minus cash held at the Bank of England, UK interbank lending and advances under sale and repurchase agreements. Two series are plotted, including and excluding “problem-country debt” (PCD) provisions. (The PCD figures were sourced from Bank of England Working Paper no. 177 The provisioning experience of the major UK banks: a small panel investigation by Darren Pain.)

In the five years from 1982 to 1986, provisions as a percentage of assets at risk totalled 4.7%. However, this understates credit impairment over the period because banks delayed taking charges for problem-country debt until 1987 and 1989. If these PCD charges are included, the 1982-86 total rises to 8.9%.

Provisions were made on a more timely basis in the early 1990s downturn and totalled 7.1% of assets at risk over 1989-93.

Based on 2008 assets at risk of £2.7 trillion, the five-year provision rates of 7.1% in the 1990s and 8.9% in the 1980s would imply aggregate credit losses for the major high-street banks of £190 billion and £240 billion respectively over 2008-12.

Such numbers are considerably larger than the Bank of England’s October estimate of £115-130 billion but this does not imply that banks require additional government financial support, for the following reasons:

1. Banks raised over £70 billion of new capital during 2008 to cover credit impairment, including £37 billion from the government.
2. The Financial Services Authority last week stated it would allow tier 1 ratios to fall to 6-7% as credit losses materialise. The average tier 1 ratio of major banks rose to more than 11% after last year’s capital-raising, implying a buffer of about £100 billion.
3. High-street banks’ net income before provisions and tax averaged £35 billion pa over 2003-07, according to the BBA. Assuming a similar run-rate before losses over 2008-12, about £175 billion will be available to shore up balance sheets and / or pay dividends. (Tax payments will be modest given credit losses.)
4. Banks’ net interest income fell to a record low 0.8% of assets in 2007, according to the BBA, versus an average of 1.3% over the previous five years. If the interest margin were restored to 1.3%, pre-tax profits would rise by about £25 billion pa or £125 billion over five years.

On reasonable assumptions, therefore, £400 billion of capital reserves and future profits could be available to set against credit losses of £190-240 billion over 2008-12. This would be sufficient to allow banks to expand their lending and even resume normal dividend distributions. Such a scenario, however, depends on the authorities affording them the flexibility to absorb losses over time, as in previous cycles. The alternative of requiring up-front accounting for future losses and further capital-raising on expensive terms is neither necessary nor desirable.