Shares digital

After nearly eight years of underperformance, chip stock valuations look tempting, with balance sheets that are cash-rich. Selective plays should yield a profit and Russ Mould surveys the semiconductor sector to find out what’s hot and what’s not for investors

by Russ Mould

Profit warnings from Sony Ericsson and American graphics chip maker NVIDIA, to name but two, have highlighted renewed weakness in key areas of chip demand, including the vital personal computer and mobile phone markets. However, balance sheets are generally strong enough and valuations low enough to have tempted trade buyers into fresh merger and acquisition activity, and the semiconductor sub-sector has already underperformed in every year bar one since 2000.

It would be unwise to be underweight, despite the risk of further earnings downgrades, as the sector should be one of the first to respond to any broader market rally. Exposure to the sector should be selective, but for those remaining die-hard technology fans who have survived the eight-year tech stock bear market, ARM Holdings still seems the most suitable vehicle.

Confidence chipped away

The technology hardware & equipment sector has been a persistent underperformer throughout 2008. It fell 34.5% up until 17 July compared with a 20.4% drop in the FTSE All-Share and ranked the sector 35th out of the 39 sector groupings which comprise the UK’s headline indices.

Much of this chronic underperformance can be pinned on the semiconductor stocks, which represents 43% of the tech sector’s £4.2 billion market cap. The 11 silicon chip-related stocks tracked by Shares have fallen by a weighted average of 49.5% since the start of 2008. Compare this with a mere 12.0% drop suffered by the leading index of American chip stocks, the Philadelphia Semiconductor Index (SOX).

The SOX’s 19 constituents include the world’s largest semiconductor manufacturer, Intel, and also the largest provider of chip-making equipment, Applied Materials. This is not to say the British chip industry does not have a position on the world stage. Under Robin Saxby, Cambridge’s ARM Holdings pioneered the fabless and chipless model whereby its processor architectures are licensed out for a fee, avoiding the costly and volatile process of chip manufacturing. CSR leads the field in Bluetooth chip designs and Wolfson has a strong position in the consumer electronics market.

The confluence of events

Such technological prowess has not stopped the UK semiconductor sector from performing poorly in 2008. This is down to a combination of macro and micro factors.

Macroeconomic: Demand for silicon chips is generally driven by consumer-focused gadgets. Rising food, fuel and mortgage bills have all begun to squeeze consumers’ ability to spend, especially as the credit crunch has crimped banks’ willingness to lend.

Gadget demand has softened, or, in the case of mobile phones, witnessed a shift away from high-end smartphones, toward cheaper versions. Alongside its first-quarter results in April Nokia warned it believes global handset revenues will decline, in euro terms, this year at least partly due to this down-shift, despite anticipated 10% unit volume growth.

Swedish rival Sony-Ericsson followed this up in June, warning of slack demand in Europe for mid and high-end phones. Some macro caution certainly coloured ARM’s outlook in February, when chief executive Warren East targeted sales growth matching at least last year’s 6%, when measured in US dollars. This disappointed analysts and drove the shares to a three-year low.

Rival semiconductor intellectual property (IP) firm, Imagination Technologies, warned in March of softening retail sales hitting demand for its PURE Digital radio business, although June’s full-year results met a lowered base of expectations.

Industry-specific: Certain industry segments, notably the commodity Dynamic Random Access Memory (DRAM) market, have been characterised by brutal pricing.

Mercifully, only memory module expert OCZ Technologies has exposure here, although plunging DRAM prices flushed out the firm’s third profit warning within 12 months back in January. A subsequent stabilisation in DRAM prices helped OCZ beat a lowered set of expectations with its full-year results in May. This offered little respite to the shares, which have since slid back to an all-time low of 17.8p. A necessary cut-back this year in capital investment in areas which have suffered from overcapacity, notably DRAM and also (non-volatile) flash memory, has taken its toll on the suppliers of semiconductor production equipment (SPE).

Bede’s shares were suspended in March after a failed bid in February and the administrators were called in a month later after the firm’s £1.1 million debt position proved too much.

Stock specific: ARM, OCZ and Imagination are not the only stocks in the sub-sector to have endured set-backs in 2008. ARC, CSR and Wolfson have each issued disappointing trading updates this year.

Semiconductor IP firm ARC trimmed back expectations in January when it warned of delays in signing a small number of new license deals and additional expenses incurred as a result of last year’s trio of acquisitions. April’s admission of further delays and some cancellations in the USA have left the shares at a four-year low of 15.8p.

CSR also languishes at a four-year low following February’s admission of market share loss to Nokia, reputedly to US rival Broadcom. Shares in Scotland’s Wolfson Microelectronics remain mired near 30-month lows, following chief executive Dave Shrigley’s admission in April his firm had lost an MP3 player design slot with an unnamed major customer, widely thought to be Apple.

Taiwanese competition, dollar weakness and a mistimed US acquisition all served to tarnish Pure Wafer, whose shares suffered an 80% one-day plunge in June when management issued a second profit warning in eight months and admitted banking covenants had been breached.

Gone are the glory days

Taking in the past ten years as a whole semiconductor stocks have been a poor investment. Intel’s share price of $20.9 compares to a share price of $20.8 exactly ten years ago, and the SOX index has not performed much better. It broke through the 400 barrier for the first time ever in early 1999 – but trades at 360.8 now, even though chip industry revenues more than doubled to $267 billion over the past decade.

This wretched performance is largely the result of a grinding de-rating of semiconductor and SPE stocks, as industry participants and analysts alike have grappled with a gradual deceleration in the industry’s long-term growth rate.

Between 1960 and 1995, the global semiconductor industry’s revenues grew at an annualised compound rate of around 16%. Over the past decade, this has slowed to just 7%. If the Semiconductor Industry Association’s (SIA) revenue growth forecasts for 2007 to 2011 are correct, compound industry revenue growth over this period will be just 6%. Even this assumption has to be under threat, since the SIA data does not factor in any period of revenue decline during 2007 to 2011. This is despite the growing likelihood of a severe economic slowdown in 2008 and 2009.

This is all a far cry from the forecasts which were being bandied about by analysts, industry experts and – worst of all – the companies themselves during the late 1990s. The cyclical peaks of 1995, 1998 and 2000 caught out those who were forecasting the chip industry would be generating $300 billion of sales first by 1998, then by 2000 and then by 2002. We still have yet to get there, as 2007’s $256 billion sales figure attests.

The overcapacity-inspired bust of 1996, the Asian crisis-induced correction of 1998 and then the bursting of the bubble in 2000 frustrated all of those forecasts, as the industry believed its own hype and built too much capacity accordingly. Even if the latest SIA forecasts are correct, industry revenues are due to exceed $300 billion in 2011, 13 years after they were initially expected to do so.

When chips were up

It would be wrong to say chip stocks have failed to generate any return since 1997. They were in the vanguard of the tech stock bubble during 1998 to 2000, when ARM’s shares soared from 43.75p to a peak of 959p and Imagination’s from 43.5p to 583.5p. A bottoming of the chip cycle and equity markets alike heralded a sharp rally in 2003/04, but since then the stocks have largely done nothing. ARM peaked out at 141.75p in 2004, ARC at 42p and Imagination at 114p. All three lie below those levels even now.

In some respects, this analysis is unduly harsh. ARM initiated dividend payments in 2004, with regard to fiscal 2003, and has since returned £280 million in cash to shareholders, including share buybacks. Imagination recorded its first annual profit since 1999 in its financial year, which has just ended.

ARC has yet to record a profit since its flotation in 2000, Bede lost money for eight straight years before the administrators were called in and IQE hopes to record its first pre-tax profit since 2003’s listing only this year.

The broad semiconductor industry’s failure to truly monitise the benign economic conditions of 2003 to 2007, when credit was cheap and consumer spending was free and easy, looks a remarkable own goal. Excess chip supply was the real culprit here, but a series of stop-start cycles, themselves the result of inventory corrections following persistent overbuilding of mobile phones, computers and other gadgets, have hardly helped.

Balance sheets offer hope

There is some good news. Firstly, the marked underperformance seen since the turn of the decade means technology has not been tainted by the ‘efficient’ balance sheet, debt-laden jiggery-pokery that has got so many other firms into trouble in the past year. Technology has already underperformed so badly over the past eight years it has to be worth a contrarian look.

Secondly, their debts may have proved too much for Bede and Surface Technology Systems but the majority of UK silicon plays are well-funded and have net cash balance sheets. Since operational gearing is high, either through owning rapidly depreciating chip fabrication facilities (fabs) or the fixed cost of vital research and development expenditure, chip and chip equipment firms tend to avoid financial gearing if they can.

ClearSpeed and Celoxica, to name but two, have tapped the equity market for extra funds in the past 12 months. Several firms, most notably ClearSpeed, remain heavily cash-consumptive and last year saw CML Microsystems cut its dividend for the second year in a row, this time to zero.

By contrast, ARM, ARC, CSR and Wolfson have £202 million in cash in the bank between them, and 11% of the sub-sector’s market capitalisation is cash. ARM looks the safest of the lot as, although its £55 million cash pile is just 5% of its market value, the Cambridge firm generates excellent free cash flow.

Thirdly, global capital investment on new fabs is expected to fall by between 15% and 20% in 2008, according to a June survey of more than 40 of the world’s leading chipmakers by broker Merrill Lynch. A planned 9% rebound in expenditure in 2009 looks modest enough and should also help demand catch up.

Better still, many of the UK semiconductor plays, such as ARM, CSR, Wolfson and Imagination do not actually make chips at all. ARM and ARC do not make silicon chips and do not even have a chip design. Instead, the so-called semiconductor IP experts license out an idea, or architecture.

Revenues come from the initial license fee, service income and royalties, once products featuring the architecture are finally sold. Imagination, CSR and Wolfson have their own designs, but again the manufacture of these are outsourced and income is generated on the basis of royalty per chips sold.

This ‘fabless’ model insulates them to a degree from the worst of any pricing collapse caused by industry overcapacity and from the savage operational gearing which results from owning a state-of-the-art fab depreciating at more than $2 million a day.

Fourthly, several years of modest sales growth have forced a fresh round of consolidation in 2008, which should help to realign supply and demand. Franco-Italian combine STMicro-electronics and NXP, which was spun out of Philips in 2006, have agreed to merge their cellular chipset activities. DRAM players Qimonda and Elpida have agreed to a joint development plan to cut costs. In the SPE arena, Applied Materials, LAM Research and KLA-Tencor have swooped for Brooks, SEZ and ICOS respectively in 2008.

Finally, this M&A activity has highlighted some attractive valuations at the remaining quoted plays.

Diodes Inc’s April £89.1 million bid for Zetex valued the Oldham-based analogue chip maker on an enterprise value (EV)/sales multiple of around one times. CSR trades at barely 0.5 EV/sales for 2008 and Wolfson 0.7, well below the multiple offered for Zetex, even though their fabless models should mean earnings and cash flow are less volatile.

Perhaps the most interesting stock on valuation grounds is Pure Wafer. The wafer reclamation expert’s debt woes make it a high-octane, high-risk stock, but its £3.5 million market valuation stands way below the historic £40.7 million cost of its factories, even when adjusted for its December net debt figure of £7.6 million.

Intel’s upbeat second-quarter numbers offered some hope, but a raft of tepid outlook statements from major global tech firms, including VMWare, Oracle and Synnex have got the technology reporting season off to a bad start. There could be more pain to come as consumer sentiment continues to weaken.

On a more positive note, an eight-year de-rating means technology stocks are nowhere near as over-hyped as they were. Value is emerging and if weak outlooks for the second half do hit the sector hard, this could be a chance to get involved.

CONCLUSIONS: SEMICONDUCTORS

Risk to earnings forecasts: 1 (5=upside risk, 1=downside risk)

Earnings predictability: 2 (5=very high, 1=very low)

Valuation: 4 (5=cheap, 1=expensive)

Balance sheet strength: 4 (5=cash rich, 1=heavily indebted)

Cashflow: 2 (5=very strong, 1=very weak)

Overlooked? 3 (5=all brokers negative, 1=all positive)

TOTAL 16 / 30

Stocks to Buy: ARM, IQE

Stocks to Avoid: ClearSpeed Technologies, Imagination Technologies

Total Broker Buy Ratings on Stocks: 24

Total Broker Hold Ratings on Stocks: 22

Total Broker Sell Ratings on Stocks: 5

SECTOR RATING - NEUTRAL

FTSE ALL-SHARE SECTOR PERFORMANCE DATA SINCE 1 JANUARY 2008

Oil Equipment & Services +4.1%

Industrial Engineering +3.2%

Electricity -0.8%

Mining -3.9%

Electrical & Electronic Equipment -5.6%

Pharmaceuticals & Biotechnology -6.8%

Chemicals -9.7%

Software & Computer Services -11.0%

Oil & Gas Producers -14.0%

General Industrials -15.2%

Equity Investment Instruments -15.9%

Gas, Water and Multi-Utilities -16.9%

Beverages -18.3%

Tobacco -19.3%

Non-life insurance -19.4%

Healthcare Equipment & Services -20.4%

FTSE All-Share Index -20.4%

Aerospace & Defence -20.6%

Support Services -21.7%

Mobile Telecommunications -22.0%

Consumer Goods -23.2%

Food & Drug Retailers -22.9%

Personal Goods -23.2%

Food Producers -23.8%

General Financial -24.9%

Fixed Line Telecommunications -25.5%

Industrial Transportation -28.3%

Automobiles & Parts -28.5%

Travel & Leisure -28.8%

Consumer Services -29.6%

Real Estate -29.8%

Media -30.6%

Life insurance -31.7%

Construction & Materials -32.3%

Household Goods -32.4%

Technology Hardware & Equipment -34.5%

Computer Hardware

Electronic Office Equipment

Semiconductors

Telecommunications Equipment

Banks -35.9%

Leisure Goods -39.2%

General Retailers -40.9%

Forestry & Paper -46.1%

Source: Datastream 17 July 2008

THE SECTOR"S PLAYS OF THE WEEK

ARM Holdings 87p BUY

Big comfort

February’s downbeat trading statement means ARM has been a disappointing performer this year, but at least chief executive Warren East’s caution established a sensible level of expectations for 2008. Further downgrades are a risk, but this year’s 31% share price swoon should largely discount this. Long-term investors should look to pick up stock if the imminent technology quarterly reporting season prompts further weakness in the shares.

A record backlog offers some comfort, as do a £55 million net cash pile and the excellent free cash flow which stems from the royalty income ARM generates each time a gadget is sold featuring one of its designs. Royalty income, which is reported one quarter in arrears, came to £32.5 million in ARM’s first quarter and was driven by the shipment of 889 million units.

A base of 542 licenses signed will continue to drive royalty income for years to come, especially as ARM’s latest generation products, the ARM11 and Cortex A9 design architectures, represented just 3% of Q1 royalty income between them.

Future cash flows are set fair, but there can be no denying the current trading environment is tough. ARM is reducing its dependence on the mobile phone market but it still generates 70% of royalty income at its PD division. The collapse in the share price of the world’s largest phone distributor, Brightpoint, suggests further estimate downgrades cannot be ruled out. Management also has to convince 2004’s acquisition Artisan, which is now the PIPD division, is back on track after a further management shake-up.

IQE 16p BUY

Funded growth

A prospective price/earnings ratio of 20.1 for 2008 is no bargain but if confidence grows in the 2009 earnings forecasts the recent share price slide down near the 12-month low of 14.75p could be an opportunity, as this leaves the stock on a PE of just ten for 2009. Last week’s trading update (17 July) certainly suggested all is still going according to plan.

Cardiff-based IQE continues to enjoy strong demand for its products from high-speed wireless communications technologies, such as third generation (3G) cellular systems, WiFi and WiMax. As a result, sales are expected to grow by 25% in 2008 and 12% in 2009, according to consensus. Better still, IQE, which prepares the wafers from which complex semiconductors made from compounds such as Gallium Arsenide are cut, is this year expected to record its first annual pre-tax profit since 2003’s listing.

IQE’s balance sheet is not typical of most semiconductor firms, in that the £62.7 million cap had £14.2 million of net debt in December, as a result of 2006’s acquisition of America’s EMD and Singapore’s MBE. Lloyds TSB was still happy enough to offer a new £15.5 million borrowing facility in January and an overseas working capital facility of £1 million provides further liquidity with which to fund growth.

ARC Intl 19.5p HOLD

Proof of life

This year’s wretched share price showing, which has seen the semiconductor IP specialist plunge from 34.5p to 19.5p, at least leaves the stock on a bombed-out valuation, from which it is hard to justify downside. The St Albans firm has yet to really prove the strategic repositioning pushed through by chief executive Carl Schlachte since his arrival in 2004 will pay off, so a watching brief is advised.

A £21.2 million cash pile provides ample support to the £24.1 million market cap, and means the £14.4 million of revenues generated by ARC last year are valued by the market at just £2.9 million. This implies an enterprise value/sales ratio for the underlying business of just 20%, on a historic basis and 16% for 2008.

There is a reason for this: ARC has yet to make a profit since its flotation in late 2000 and further losses are expected for 2008 and 2009. A quartet of acquisitions, Teja, Tenison, Alarity and Sonic Focus have seen ARC position itself in the multimedia IP areas, but license wins have been slow to come and royalty income remains years away.

A private equity approach cannot be ruled out completely, given the lowly valuation, but ARC’s lack of a proven history of cash generation would suggest a bid approach is unlikely, at least in the current testing industry environment.

Wolfson Microelectronics 110p HOLD

Don’t touch that dial

Chief executive Dave Shrigley’s view is strong demand from the mobile phone and automotive markets will help compensate for any second-half shortfall from portable music players. If he is correct Wolfson’s shares are very cheap. However, an unhelpful macro-economic environment suggests both Shrigley, and consensus earnings forecasts calling for an 11% drop in earnings per share this year and a 14% rebound in 2009, are too optimistic.

March’s admission the consumer electronics chip designer had lost a key MP3 design slot with a major (unnamed) customer, widely assumed to be Apple, has hammered Wolfson’s share price this year and left the stock hovering near its autumn 2004 all-time low of 100p. At first glance this has left the shares attractively valued on a prospective price/earnings multiple of 8.8 times for 2008 and 7.7 times for 2009.

All the available evidence points to a deteriorating mobile phone market. The automotive sector’s woes are abundantly clear to all, judging by the share prices of dealers such as Pendragon and manufacturers Porsche and GM. This means earnings estimates are likely to prove too optimistic, although the marked underperformance has at least priced in some of the likely disappointment. At least an enterprise value/sales ratio of 0.7 is not expensive and will provide support in the event of any further earnings set-backs.

ClearSpeed Technology 16.5p SELL

Jam tomorrow

Even though shares in ClearSpeed have plunged from 2005’s all-time high of 290p to just 16.5p, and the firm’s technology is rich with promise, the shares are still to be avoided as it is the sort of ‘jam tomorrow’ stock which is likely to do poorly in this equity market environment. A cash pile of £14 million exceeds the processor chip designer’s market capitalisation of £10.3 million, but analyst forecasts assume aggregate pre-tax losses totalling £14 million across 2008 and 2009.

Progress toward a maiden profit has proved slow, even though its new CSX700 processor has already been approved by hardware giant Hewlett Packard for use in its blade servers and by BAE Systems for future US government satellite processing requirements. Projected sales are £1.2 million in 2008 and £2 million in 2009. These forecasts cannot be taken for granted, despite a trading statement earlier this month, which revealed interim sales doubled to £438,000.

Management is aware the overall environment is getting tougher, judging by the announcement of a restructuring plan, designed to generate £3 million in annualised cost savings.

Imagination Technology 66p SELL

Not looking pretty

In its fiscal year which ended in April, Imagination recorded its first annual pre-tax profit since 2001. Its last annual profit coincided with the peak of the last chip cycle and, having invested heavily in the intervening seven years, the Hertfordshire firm could again be unlucky enough to find an industry downturn thwarting its ambitions.

Imagination has already issued one lukewarm trading statement this year, back in March when the graphics IP and digital radio leader noted weak demand for PURE radio products. The shares may have jumped 37% in two days after June’s full-year numbers, but earnings estimates and the stock had both collapsed following March’s commentary so expectations had been dramatically lowered.

Last year did see a 46% rise in licensing income and 24% growth in royalties, but the numbers did not impress upon closer inspection. A seven-month second half still saw pre-tax income of just £0.6 million, against £1.3 million in the first six months. There is great long-term potential, as only 26 of Imagination’s 67 design wins are in production. A short-term loss of momentum could easily leave consensus forecasts of a doubling earnings per share to 2.0p in the next two years exposed on the downside.

RISING STAR

ARM readies its muscles

The pre-eminent chip firm has a depressed sector to face but will come up trumps

by Russ Mould

It may seem odd to identify a firm which has just celebrated its tenth anniversary as a public company as a rising star, but ARM is the only stock of the 11-strong sub-sector which is expected to increase its profits and pay a dividend in 2008.

To ignore its pre-eminence within the sector would therefore be to do its management team a disservice, and if the sector does come back into fashion, the Cambridge firm will be one of the market’s first ports of call. A prospective price/earnings ratio (PE) of 19.1 is not bargain-basement territory but the stock’s chart appears to have found a base near 80p and a near halving of the share price over the past year means a lot of bad news is already priced in.

A net cash balance sheet, regular dividend payments and a share buyback programme make ARM easily the most financially robust of all of the UK semiconductor sector. Since 2004, the £1.1 billion cap has returned £280 million to shareholders in cash and the company’s financial strength stems from its licensing model, which it pioneered in the 1980s and early 1990s. The company does not get involved in the difficult and expensive process of silicon chip manufacturing. Instead, it licenses out its processor architectures for a fee and then collects a royalty each time a chip featuring its designs is manufactured.

License for profit

Much of ARM’s licensing income is inherently defensive in nature, as it is driven by customers’ need to make research and development spending more efficient, so new chip products can be brought to market as quickly as possible. Licensees also get access to the Cambridge firm’s intellectual property (IP) expertise in low-power, high-performance processing, which has made ARM’s designs so ideally suited to products where portability and battery life are vital differentiators.

That is not to say licensing income is immune from cuts when industry conditions get really tough – ARM signed just eight licenses in Q3 2002, down from 27 in the prior quarter, as the global semiconductor industry really hit the buffers – but it should offer some downside protection.

A base of 542 licenses sold will continue to drive royalty payments, which averaged 6.2 cents per unit in the first quarter, for years to come. ARM is positioned to cash in on long-term growth in demand for mobile phones, set-top boxes, printers, cars, and telecoms equipment.

A market to convince

If the long-term picture is strong, the short-term is undeniably more difficult to call. Weakness in the US dollar, the potential challenge from Intel’s Atom chip, an ongoing failure to convince the market 2004’s acquisition of Artisan was not a $913 million mistake have all weighed on sentiment this year.

Worries about Atom seem overdone. Intel’s last foray into mobile communications was a disaster and ended with the company selling its assets to rival firm Marvell in 2006. The same problems that prevented a breakthrough then still seem relevant today.

Firstly, the mobile phone makers do not wish to leave themselves exposed to one dominant chip supplier and software vendor as the PC market has done, so Intel will encounter considerable customer resistance. Secondly, Intel’s microprocessing prowess is based on the PC market and as a result its designs have not been as well suited as ARM’s to the mobile phone market, since they are much more power hungry.

Atom, and its successors, will be less power consumptive, but low power requirements (and therefore high battery life) has been one of ARM’s long-running strengths. At the firm’s May analyst meeting, ARM demonstrated how its Cortex A9 processor design will consume over a quarter less power than the rival Intel offering.

A tough trading environment, which ARM was one of the first chip firms to flag back in February, has been a further handicap. An outlook statement, which accompanied the full-year 2007 results, guided for sales growth at least matching last year’s 6%, when measured in US dollars.

If ARM can meet this goal for the year, the shares should rebound strongly, as recent share price weakness means the stock is pricing in further downgrades and in the current environment any firm which can generate organic secular growth will become highly prized.

Interim results on 30 July will be the next test to see whether ARM can deliver on its long-term goal of 15% to 20% compound growth per annum in royalties and a single-digit compound advance in licensing.

CONCLUSIONS: ARM HOLDINGS

Risk to earnings forecasts: 2 (5=upside risk, 1=downside risk)

Earnings predictability: 3 (5=very high, 1=very low)

Valuation: 4 (5=cheap, 1=expensive)

Balance sheet strength: 5 (5=cash rich, 1=heavily indebted)

Cashflow: 4 (5=very strong, 1=very weak)

Overlooked? 3 (5=all brokers negative, 1=all positive)

TOTAL 20 / 30

RATING – BUY

When the chips are down

Weakness has hit the semiconductor sector once again and there is plenty to suggest it has further to drop

by Simon Griffin

The chart of the UK semiconductor sub-sector clearly tells a story. From the heady highs seen at the top of the technology bubble in 2000, the sector (drawn here on a capitalisation-weighted basis) declined over 90% to the lows seen in the autumn of 2002 and early 2003.

A bull run in the broader market from 2003 to 2007 helped the grouping rally 236% from its lows but even then it did not manage to recover even a quarter of the ground it had previously lost.

Worse, the chip stocks peaked in May 2006, a year ahead of the market as a whole. Weakness has since set in once more. The trend line that had supported the preceding bull run was broken when the index, having failed to supplant the 2006 high, headed south in the autumn of last year. This signalled the sector was likely to fall significantly and we have now seen a crunching 65% drop from the 2006 highs.

The proximity of the previous grand lows suggests the market could be reaching levels that might turn the trend once more, although we could still be in for a further drop of 14% even if this is so.

The worry is the correction seen between March and May of this year could be interpreted as a bear flag pattern, which would point to continuing losses well below the 2002 low. Currently the trend is down and we would need to see a hefty 25% climb from current levels to postulate the end of the selling and the possible start of a fresh bull trend.

ARM Holdings (ARM)

BUY - 87p

TARGET - 144p

STOP LOSS - 78p

The ARM chart mirrors that of its sector, at least in the dramatic sell-off between 2000 and late 2002. Though it too recovered in the early stages of the 2003 to 2007 bull market, the shares then stalled near 144p and have since found life hard above this level. The shallow bull trend line drawn off the lows seen in June 2003 and July 2006, which briefly held last December, gave way as the year turned and the shares dropped to test support from the low of 81p seen in September 2004.

Following a recovery move to test resistance from its 200-day average in mid-May, a second test of this support level has taken place and suggests this marks a low point for the shares. If we subscribe to the ranging sideways stance suggested by the 144p to 81p area, then a speculative buy at current levels looking for a recovery toward the former could prove successful.

The first target would be the 200-day average, which is still declining and currently close to 110p. One thing worth bearing in mind is the eventual breakout from a range will produce a move proportionate to the length of time the shares have been within the range. At some stage ARM will therefore make a dramatic move but if investors respect the range boundaries they should be on the right side of it when that day comes.

Wolfson Microelectronics (WLF)

BUY - 110p

TARGET - 180p

STOP LOSS - 97p

Having only come to the market in late 2003, the shares have been quite a roller-coaster ride for holders. Initially they dipped well below their offer level and after 12 months were down some 60%. This type of post-launch behaviour usually makes any recovery a slow and torturous affair.

In this case Wolfson’s shares took off and over the next two years provided a return of 460% before they hit a high of 556p just as the sector itself peaked. Subsequently we have seen yet another large move, this time back down to test the late 2004 lows close to 100p. Could investors have traded these moves?

The shares certainly seem to respect the 200-day average and trading off this line would have captured the lion’s share of the profits. With the shares now on their all-time lows, they seem to be offering the opportunity for the adventurous to trade.

Any bounce would be expected to attack initial resistance from previous price congestion close to 140p. If this level, which is likely shortly to coincide with the 200-day moving average, can be broken above then there will be a double bottom in place. This would then suggest a move to 180p, coincidently close to the level of the descending trend line which has consistently called the tops of corrections during the sell-off.