Blog

Does the economy face a temporary period of stagnation or mild contraction to be followed by a recovery from the first half of 2009, or a full-scale recession of the sort experienced in the mid 1970s, early 1980s and early 1990s, involving a fall of several percentage points in GDP over a period of at least a year?. These notes have argued in favour of the former scenario for three main reasons: the interest-servicing burden on households has increased less sharply than before prior recessions; monetary growth has weakened to a lesser extent, at least to date; and the large fall in the sterling exchange rate over the past year should offer external support to UK activity.

Recent extraordinary financial events have clearly increased the risk of a full recession. In particular, financial turmoil may cause consumers and businesses to postpone spending decisions, possibly initiating a self-feeding economic downturn. However, the judgement here is that the less malign scenario remains the more likely, albeit involving a bumpier ride than previously envisaged. As explained further below, not all recent developments have been discouraging. The latest trade figures support optimism that net exports will cushion domestic demand weakness. Annual retail price inflation should fall precipitously over the next 12 months, providing a welcome boost to spending power. Meanwhile, financial events have created scope for the MPC to cut official rates earlier and by more than was previously likely without damaging its inflation-fighting credibility.

An improving trade account has provided major support to the US economy, with net exports contributing 1.7 percentage points to GDP growth of 2.2% in the year to the second quarter. The consensus has been sceptical of a comparable trade boost to UK activity, despite sterling’s plunge and the much higher shares of exports and imports in GDP. Independent forecasters expect net exports to contribute 0.4pp to GDP growth in calendar 2008 and 0.6pp in 2009, according to the Treasury’s monthly survey. Recent figures suggest a larger boost. Export volumes of goods excluding oil and erratic items were 3.6% above their 2007 average in June / July versus a rise of just 0.2% in imports. Suppose total real exports and imports, including services, grow at these rates for the year as a whole – plausible given that the full benefit of sterling depreciation has yet to be felt. Net trade would then add 0.9pp to calendar 2008 GDP expansion.

While the credit crisis is more often blamed, a squeeze on real income and money supply growth caused by this year’s surge in inflation has also contributed significantly to current economic weakness. The potential benefit from a likely reversal of this spike should therefore not be underestimated. The chart shows forecasts for annual CPI and RPI inflation through to the end of 2009 based on the following assumptions: 1) unprocessed food price inflation slows gradually from 13% in August to 2.5% by December 2009; 2) electricity / gas tariffs rise by a further 10% in September but energy prices are stable thereafter; 3) “core” CPI inflation – excluding unprocessed food and energy – slows gradually from recent rates, reflecting economic weakness and a fading impact from sterling’s decline; 4) the housing depreciation component of the RPI, which tracks house prices, falls by 15% by mid 2009, stabilising thereafter; and 5) Bank rate is cut by half a point to 4.5% by the end of 2008 (see below), remaining stable in 2009. On these assumptions, annual CPI inflation returns to the 2% target in September 2009 while RPI inflation falls below 2% in June next year, reaching 1.1% in September. Of course, further official rate cuts during 2009 would imply still-lower RPI numbers. While interest rate reductions contribute, falling house prices are the key reason for the forecast RPI / CPI divergence. The housing depreciation component carries a 5.5% weight in the RPI so the assumed 15% annual decline in mid 2009 subtracts 0.8 percentage points from the annual RPI increase.

Minutes of the September MPC meeting partially vindicate the recent dovish shift in the MPC-ometer model described in previous monthly notes. The vote changed from 1-7-1 in August (Besley seeking a 25bp hike, Blanchflower a 25bp cut) to 8-1 (Blanchflower voting for a 50bp cut). Taking into account the September outcome and available data on the other inputs, the model suggests a knife-edge October decision. Further weakness in consumer and business surveys released in late September and early October – plausible in light of recent financial events – would tip the balance in favour of a cut. Market developments will also be important. MPC members were concerned by the plunge in sterling during August but the effective index is currently 3% higher than at the time of the September meeting. Meanwhile, recent rises in unsecured interbank lending rates, unless rapidly reversed, threaten renewed upward pressure on household and corporate borrowing costs. As conventionally used, the MPC-ometer attempts to answer the question of whether a change in Bank rate is warranted by incoming economic and financial data. If the target is changed from Bank rate to three-month LIBOR, currently 6.1%, the model suggests a near-certain 25bp October cut and a further 50bp reduction by year-end, based on plausible assumptions about the inputs. Against these arguments, the MPC may feel itself constrained by a further rise in annual CPI inflation in September, possibly to 5% or above, as well as rapidly deteriorating fiscal prospects, which could undermine confidence in the wider macroeconomic policy framework. However, the disinflationary shock implied by recent financial events should allow the Committee to play down the former, while fiscal profligacy may affect the extent of the peak-to-trough decline in rates rather than the timing of the first cut.

Some commentators have suggested the Paulson / Bernanke financial rescue plan represents a “monetisation” of illiquid mortgage-backed securities, implying longer-term inflationary consequences. On current information, such concerns appear unwarranted.

“Monetisation” would involve one or both of the following:

• An increase in the monetary base, i.e. currency in circulation and banks’ reserves held at the Fed.

• An increase in the broad money supply, i.e. cash, deposits and other liquid assets held by the non-bank private sector.

The Fed has hugely expanded its lending to the banking system against lower-quality collateral over the last 12 months but has sterilised the impact on the monetary base by selling Treasury securities. There is no current reason to think this approach will change. The further measures announced in recent days – including the loan to AIG, a new facility allowing banks to borrow to buy asset-backed commercial paper and planned Fed purchases of agency securities – will be financed mainly by the Treasury issuing additional bills and depositing the proceeds with the Fed, implying no impact on bank reserves.

Under current arrangements whereby the Fed does not pay interest on bank reserves, failure to sterilise the impact of its lending on the monetary base would push the effective fed funds rate down to zero. To keep the rate near the policy target of 2% while expanding the monetary base, the Fed would have to start paying interest on reserves at close to this level. There has yet to be any discussion of this possibility.

With respect to the broad money supply, the key point is that the Paulson / Bernanke plan involves the government buying suspect assets from banks rather than non-banks so there is no direct impact on the money holdings of the latter. If the scheme is financed by selling additional Treasury bills to banks, the net effect will be to increase the proportion of high-quality liquid securities on their balance sheets while leaving total assets unchanged. To the extent that funds are raised by selling additional Treasury securities to the non-bank private sector, there will actually be a negative first-round impact on the broad money supply and aggregate bank assets.

Of course, a successful scheme resulting in a normalisation of money and credit markets would increase banks’ willingness to lend, implying an indirect boost to monetary expansion.

Should the US authorities consider measures to boost broad money growth as part of their efforts to restore financial stability and stave off economic weakness? My favoured broad liquidity measure – M2 plus institutional money funds, commercial bank large time deposits and commercial paper outstanding – is still rising at a 9% annual rate, suggesting concern over monetary deficiency is premature. However, the shorter-term trend is weaker and bears close monitoring – see chart.

The Resolution Trust Corporation (RTC) was set up in 1989 to manage and dispose of the assets of failed savings and loan (thrift) institutions. According to a General Accounting Office report, the RTC closed 747 institutions with $402 billion in book value of assets. By the time it was wound up in December 1995, the RTC’s remaining assets were down to $8 billion and it had incurred estimated direct losses of $88 billion.

The Treasury, Federal Reserve and Congressional leaders are reportedly discussing plans to create an RTC-type body to purchase so-called failed assets from US financial institutions. Such an approach has been endorsed by a former Fed chairman, a former Treasury secretary and a former US comptroller of the currency. The original RTC, however, assumed assets only after savings institutions had failed and been placed into conservatorship (the current status of Fannie Mae and Freddie Mac) or receivership. By contrast, the latest proposal is designed to prevent further financial failures.

The original RTC bailed out the depositors of failing institutions but not equity-holders or junior creditors. Market participants are betting that the new scheme will be different, contributing – together with the ban on new short sales – to today’s surge in financial shares. The benefit to equity-owners, however, will depend on the price at which any new RTC-type body purchases toxic assets. Who will establish the “correct” level, and how? Too high a price will enflame criticisms of a bail-out and bolster calls for tighter regulation. Too low a price risks leaving banks with insufficient capital to weather future financial storms.

The Fed has hugely increased its lending against lower-quality collateral since the credit crisis broke. It has, however, sterilised the impact of this lending on the volume of cash circulating in the banking system by selling Treasury securities. This was necessary to prevent the fed funds rate from falling beneath its policy target.

The change in the composition of the Fed’s assets has raised concerns about its financial strength. Its non-conventional lending will rise significantly further as a result of the $85 billion credit facility for AIG agreed this week. Sterilising this liquidity injection via further sales of Treasuries would reduce its holdings of such securities to undesirably low levels.

The US authorities therefore yesterday announced a new “supplementary financing programme”, under which the Treasury will sell additional Treasury bills and deposit the proceeds at the Fed. This has the effect of draining liquidity from the market without reducing the Fed’s own holdings of Treasuries further.

Should the Fed follow the example of the Bank of Japan in 2001 and stop sterilising its liquidity injections? The BoJ’s policy of “quantitative easing” flooded the banking system with reserves – see chart – and is argued by some to have been instrumental in Japan’s escape from deflation.

Under current arrangements whereby the Fed does not pay interest on bank reserves, unsterilised liquidity injections would push the Fed funds rate down to zero. The Fed could, however, avoid such an outcome by paying interest on bank balances at close to the fed funds target rate, currently 2%.

The main objection to such a suggestion is that, unlike Japan in 2001, the US still faces inflationary rather than deflationary risks. Flooding the banking system with reserves would risk undermining the dollar and reigniting commodity prices. The surge in the gold price over the last 24 hours is a taster of the possible implications.

Minutes of the September MPC meeting partially vindicate the recent dovish shift in my MPC-ometer. The vote changed from 1-7-1 in August (Besley seeking a 25 bp hike, Blanchflower a 25 bp cut) to 8-1 (Blanchflower voting for a 50 bp cut).

Taking into account the September vote and available data on the other inputs, the model suggests a knife-edge October decision. Further weakness in consumer and business surveys released in late September and early October – plausible in light of recent financial events – would tip the balance in favour of a cut.

Exchange rate developments will also be important. MPC members were concerned by the plunge in sterling during August but the effective index is currently 1.5% higher than at the time of the September meeting.

The MPC will wish to avoid the impression of cutting rates to rescue miscreant banks but current financial turmoil clearly has negative implications for credit supply and the wider economy.

Labour market figures also released today showed a shock 32,500 rise in claimant-count unemployment in August, boosting fears that the economy is already in recession. However, my research suggests that job vacancies are a better coincident indicator than unemployment. The 8% drop over the last three months is consistent with stagnant rather than contracting GDP – see chart.

The annual increase in UK consumer prices climbed further to 4.7% in August but retail price inflation slowed from 5.0% to 4.8%. The RPI measure will fall beneath CPI inflation over coming months. The gap between the two may exceed one percentage point by mid 2009 – the largest since 1993.

The chart shows my updated forecasts incorporating the latest numbers. Key assumptions include:

• Unprocessed food price inflation slows gradually from 13% in August to 2.5% by December 2009.

• Electricity / gas tariffs rise by a further 10% in September but energy prices are stable thereafter.

• “Core” CPI inflation – excluding unprocessed food and energy – slows gradually from recent rates, reflecting economic weakness and a fading impact from sterling’s depreciation.

• The housing depreciation component of the RPI, which tracks house prices, falls by 15% by mid 2009, stabilising thereafter. (The CPI excludes housing depreciation.)

• Bank rate is cut by half a point to 4.5% by the end of 2008, remaining stable in 2009.

On these assumptions, annual CPI inflation returns to the 2% target in September 2009. However, RPI inflation falls below 2% in June next year, reaching 1.1% in September. Of course, further Bank rate cuts during 2009 would imply still-lower RPI numbers.

While lower interest rates contribute, falling house prices are the key reason for the forecast RPI / CPI divergence. The housing depreciation component carries a 5.5% weight in the RPI so the assumed 15% annual decline in mid 2009 subtracts 0.8 percentage points from the annual RPI increase.

The coming plunge in RPI inflation will lower inflation expectations, subdue wage settlements and boost real income growth. It should also make it easier for the MPC to justify cutting Bank rate while CPI inflation is still above target. I expect the first move to occur as early as next month.

Last week the US authorities rescued Fannie Mae and Freddie Mac, judging them “too big to fail”. This week the authorities have refused to bail out Lehman, on the basis that the firm’s failure is unlikely to trigger a systemic crisis. These judgements are defensible.

The risk of Merrill Lynch being the next domino to fall has been removed by its purchase (at a premium) by Bank of America. With the Fed ready to supply unlimited liquidity against an expanded range of collateral, including equities, fears of financial Armageddon are overdone.

Credit conditions will remain tighter for longer, with negative implications for the economy. However, lower energy prices will provide some offset – global headline inflation is about to fall sharply, boosting real incomes and creating scope for central banks to cut interest rates.

Based on recent crude oil and natural gas prices, the annual change in the energy component of the US consumer price index may fall from +29% in July (August numbers are released tomorrow) to -10% in early 2009 – see chart. Assuming no change in non-energy inflation, this would push headline CPI inflation down from 5.6% in July to about 2% by early next year.

In the UK, CPI inflation is likely to peak at about 4.8% in September before embarking on a steady decline. With wage settlements stable, money growth weakening and the economy stagnant, the MPC should be able to cut interest rates either next month or by November at the latest.

An earlier post argued that UK house prices needed to fall by a further 6% from their July level to bring the rental yield on housing back to its long-term average, assuming stable rents. If the rental yield were to overshoot the average to the same extent that it undershot in 2007, a decline of 21% would be necessary. These could be considered best and worst case scenarios for house prices.

The two scenarios would imply peak-to-trough falls in house prices of 16% and 30% respectively.

Prices dropped by a further 1.8% in August, according to the Halifax index. The required additional declines are therefore now 4% based on the long-term average yield and 19% in the yield overshoot scenario.

The chart below compares the scenarios with the last three house price busts – 1973-77, 1980-82 and 1989-96. The comparison is in real terms – relative to the retail prices index – because general inflation carried a greater burden of valuation adjustment in the prior episodes.

The scenario paths assume that remaining house price adjustment occurs smoothly over two years while the RPI rises by 3% per annum.

In the best case scenario the peak-to-trough decline in real house prices would be much larger than in 1980-82 but less severe than in 1973-77 and 1989-96. The worst case scenario would imply greater damage even than in 1973-77.

Restricted mortgage availability and coming labour market weakness clearly suggest an outcome closer to the worst case. However, prices are now falling faster than at the comparable stage of previous busts. If the decline continues at its recent pace, nominal prices could be nearing a trough by next spring.

An improving trade account has provided major support to the US economy. Net exports contributed 1.7 percentage points to GDP growth of 2.2% in the year to the second quarter.

The consensus is sceptical of a similar positive impact from trade in the UK, despite sterling’s plunge. Independent forecasters expect net exports to contribute 0.3 percentage points to GDP growth in calendar 2008 and 0.4 pp in 2009, according to the Treasury’s monthly survey.

Recent trade figures suggest a larger boost. According to the July release issued this morning, export volumes of goods excluding oil and erratic items were 3.6% above their 2007 average in June / July versus a rise of just 0.2% in imports. Suppose total real exports and imports – including services – grow at these rates for the year as a whole. Net trade would then add 0.9 percentage points to calendar 2008 GDP expansion.

The ratio of export to import volumes in July was the highest since 2006 – see chart.

Public sector net borrowing totalled £35 billion in 2007/08, equivalent to 2.5% of GDP. The average forecast of economists surveyed by the Treasury is for a rise to £47 billion this year and £50 billion in 2009/10, or about 3 ¼% of GDP in both years. I think the consensus is underestimating the sensitivity of the public finances to the economic cycle. Borrowing could reach about 3 ½% of GDP this year and over 5% in 2009/10, implying nominal figures of £50-55 billion and over £75 billion respectively.

Conceptually, the public sector balance can be split into discretionary and non-discretionary components. The analysis below assumes the government chooses the percentages of GDP accounted for by spending on goods and services and indirect tax revenues. Other elements of the public finances – mainly benefit spending and direct tax receipts – are defined as non-discretionary, i.e. determined by the economic cycle, financial market developments and other factors. (Of course, policy decisions also affect these elements but their impact on year-to-year changes is small relative to cyclical influences.)

The first chart shows annual changes in the non-discretionary public sector balance, defined as above and expressed as a percentage of GDP, together with the fitted values of a model based on current and lagged rates of change of GDP and stock prices. (The labels on the horizontal axis refer to financial years ending in March of the year stated.) Despite its simplicity, the model is able to explain the major swings in the non-discretionary balance over the last 20 years.

Using the model to forecast requires assumptions about GDP and stock prices. GDP is projected to be unchanged between the second quarter of 2008 and the first quarter of 2009 and to grow at an annualised rate of 2% thereafter – close to the current consensus forecast. Similarly, the FT all-share index is assumed to be stable at its current level until the first quarter of next year and then to rise by 10% per annum. Based on these inputs, the model projects a year-over-year deterioration in the non-discretionary balance of 1.1% of GDP in 2008/09 and 1.7% in 2009/10.

The change in the overall public sector balance will also depend on the discretionary component. For simplicity, spending on goods and services and indirect tax revenues are assumed to remain stable as percentages of GDP in 2008/09 and 2009/10. (The former assumption, in particular, may be optimistic given a possible pick-up in public sector wage settlements.)

The second chart shows the overall balance as a percentage of GDP, including the forecasts for 2008/09 and 2009/10. Projected borrowing of 5.3% of GDP in the latter year would be the highest since 1995/96. In a full recession – not the assumption here – borrowing would probably challenge the 7.8% of GDP peak reached in 1994/95.