Vanguard analysis finds that oil price spikes don’t lead to lower stock returns; Federated and Barclays like Mexico; Morgan Stanley finds that ETFs are tracking their markets more accurately; Keefe, Bruyette & Woods report that demand for “life-cycle” type investment products should grow.
Joseph H. Davis, Roger Aliaga-Diaz, Vanguard Investment Counseling & Research
Oil, The Economy and the Stock Market
In the past, sharp rises in oil prices have coincided with bear equity markets, sparked higher inflation and government bond yields, and presaged economic recessions. Recently, as oil prices have leapt past $100 a barrel, some analysts warn of a prolonged period of “stagflation,” or persistent inflation coupled with high unemployment rates. Other analysts are more sanguine, arguing that demand-driven oil-price shocks are less relevant for modern economies that have become less energy-intensive.
This paper addresses several related questions. First, what impact does a given oil-price shock have on economic growth, and have those effects changed over time? How should monetary policy respond to oil-price shocks, now and in the future? How has the stock market—across countries and sectors—reacted to changes in oil prices? Does this response vary depending upon the direction or source of oil-price movements? Finally, do fundamentals explain the rapid rise in oil prices to more than $100 a barrel? We focus our analysis on the U.S. economy and financial markets, although our results are also applicable to the U.K. and broader European economies.
We find that the negative short-run impacts of oil-price shocks on U.S. economic growth since the early 1980s are similar to those in earlier decades. Indeed, oil-price shocks can still induce a U.S. recession, given the inherent lag between a change in monetary policy and its impact on real growth. However, the negative “second-round” real GDP effects of oil-price shocks witnessed during the 1970s have been notably absent over the past 25 years. Our empirical analysis attributes this phenomenon primarily to the more muted response of long-term inflation expectations and long-term interest rates to oil-price movements. This change enhances the present-day Federal Reserve Board’s flexibility in addressing oil-price shocks that, in the past, presented a difficult policy trade-off with respect to stagflation. Our simulations indicate that today it would be appropriate for the Federal Reserve to lower short-term interest rates in response to an oil-price shock, as long as long-term inflation expectations do not rise from their desired level. We provide rules-of-thumb for determining the proper change in the Federal funds rate for a given oil-price shock under alternative macroeconomic scenarios.
We find that global stock markets have responded fairly symmetrically and inversely to changes in oil prices since the 1970s, with a 10% increase in oil prices associated with a statistically significant 1.5% lower total return. However, when we decompose movements in oil prices, we find that the stock market’s reaction to oil-price increases varies dramatically, depending on the source of the oil-price shock. Specifically, the stock market—particularly in the U.S. industrials and materials sectors—responds quite favorably to oil-price increases attributed to global-demand shocks. A key implication of these stock regressions is that oil-price increases do not uniformly lead to lower stock returns. Interestingly, our oil-price decomposition suggests that the recent surge cannot be fully explained by oil-supply disruptions, global-demand fundamentals, or the decline in the value of the U.S. dollar. Do higher oil prices matter?
Throughout the 1990s, crude oil prices in the United States averaged roughly $20 a barrel, with highs of $35 a barrel set during the 1990–1991 Persian Gulf War and lows of approximately $11 a barrel reached after the 1998 emerging-market crises.2 Since that time, petroleum prices have skyrocketed to once unimaginable levels. In early 2008, oil prices topped $110 a barrel, breaking the inflation-adjusted highs established in April 1980. As oil spot prices have risen, so too has the futures market’s perception of oil’s long-run or “fair” value. In the past, dramatic oil-price increases have often been caused by oil-supply disruptions centered in the Middle East. Notable oil-supply shocks include the 1956 Suez Crisis, the 1973 Arab–Israel War, the beginning of the Iranian Revolution in 1978, the onset of the Iran–Iraq War in 1980, and the 1990 Persian Gulf War. Today, some analysts point out that the dramatic rise in oil prices from $20 a barrel in 2001 to more than $100 in 2008 is primarily the result of the world’s first global-demand shock.
Throughout this decade, the global economy has expanded robustly, and commodity demand has surged. The emerging-market economies of China and India have more than doubled their combined share of world oil consumption since 1990, and have accounted for more than one-third of consumption growth since 2000.3 Many experts anticipate that China’s demand for oil will double again in the next two decades. In the past, sharp rises in oil prices have sparked higher inflation. More recently, as oil prices rapidly surpassed $100 a barrel, some analysts have warned of a prolonged period of U.S. stagflation, with the economy’s inflation and unemployment rates rising in tandem. Economists have long observed that oil-price shocks tend to precede U.S. recessions, as consumer discretionary spending declines and future investment plans are postponed.4 To demonstrate the powerful effect that oil-price shocks can have on the business cycle, we compare the year-over-year growth in actual U.S. real GDP with what it would have been under the counterfactual case of no oil price increases.
According to this econometric exercise, the U.S. economy would have avoided five out of the past six recessions if oil prices had remained unchanged.5 In this hypothetical world, the mild 1990 and 2001 recessions would not have been recessions at all, but rather periods of below-trend growth. This supports the work of Hamilton (1996, 2003) and others, who argue that spikes in oil prices are the primary cause of every U.S. recession since World War II.
The results would suggest to some that the recent marked rise in energy prices portends a prolonged period of economic weakness. Of course, other analysts are more sanguine, arguing that oil-price shocks are less relevant for modern economies that have become less energy-intensive and even stagflation-proof. Many industrialized countries now use less energy per unit of output, thanks to technological advances, a more service oriented economy, and improved living standards. The U.S. economy—the world’s largest consumer of oil—is 40% less oil-intensive per unit of output than it was in 1980. While petroleum remains a prominent input for the transportation industry, other sectors have become less oil-dependent since the 1970s. A good example is the electric utility industry. According to a 2004 report by the Energy Information Administration (EIA), today oil accounts for less than 2% of electric utility generation in the United States, compared with approximately 17% in the early 1980s. During that period, living standards and real wages increased for many U.S. households. Today American consumers dedicate approximately 6% of their personal expenditures to energy goods and services, down from a high of more than 9% in the early 1980s, but up from its nadir of 4% in 1999.
Our analysis showed that while the “first-round” impact of oil-price shocks on U.S. economic growth has not changed materially over time, their formerly negative “second-round” effects have been notably absent over the past 25 years, as illustrated by oil’s nearzero impact on long-term inflation expectations. Since oil-price shocks now represent a lessstagflationary policy trade-off, we concluded that the Federal Reserve should lower short-term interest rates in response to an oil-price shock under certain (but not all) macroeconomic scenarios. For domestic and international stocks, our simple regressions revealed the anticipated inverse relationship, with a 10% increase in oil prices associated with a statistically significant 1.5% lower total return. However, we have also shown that the stock market’s reaction varies dramatically, depending on the source of the oil-price shock, and that global stocks—particularly in the industrials and materials sectors—respond quite favorably to oil-price increases attributed to global-demand shocks. A key implication is that oil-price increases do not uniformly lead to lower stock returns. Interestingly, our oil-price decomposition suggests that oil’s recent surge beyond $90 per barrel cannot be explained by supply disruptions, global demand fundamentals, or the depreciation of the U.S. dollar.
Audrey H. Kaplan, Federated Investors
When it comes to Latin America, we're singing, 'Oh, Mexico'
It's unfortunate, though understandable, that much of the attention Mexico attracts these days is focused on drug cartels and murders. But the violence is overshadowing what we believe to be a compelling investment story. Low costs, above-trend economic growth and a buoyant manufacturing sector are among the reasons our neighbor south of the border has emerged as the lone overweight Latin American country in our international portfolios.
Mexico's ascent comes as we've reshuffled the deck a bit among our favorite emerging-market countries. Overall, the emerging markets remain a solid long-term theme, with above-average growth prospects, comparatively sound fiscal balance sheets and improving internal demand. But we've pared some of favorites in recent months, including Brazil, on rising inflation and monetary tightening. Yet we still see near-term value in select countries, notably South Korea in Southeast Asia and, in Latin America, Mexico.
Mexico's central bank this month raised its real GDP growth forecast for 2011 to 3.8% to 4.8%, up from a previous range of 3.2% to 4.2%, and estimated growth for all of 2010 will come in at 5.4% once the final tally is done. Manufacturing is accelerating (manufacturing output jumped from a robust 4.6% annual pace in October to a rollicking 6.6% rate in November); domestic demand is strengthening (the latest investment and auto sales reports show Mexican businesses and consumers continue to pick up the pace of spending); and the outlook for its major trading partner, the United States, is on the rise (GDP forecasts for the U.S. have been bumped up significantly since the end of last year, with the Blue Chip Economic Indicators consensus for 2011 now at 3.5%. At Federated, we're forecasting 3.6% growth).
At the same time, the inflationary pressures that are sparking rate increases and other tightening measures in many emerging-market countries have yet to rear their head in Mexico. In fact, the pace of price increases in Mexico actually fell in January from December, prompting the Bank of Mexico to lower its CPI forecast for 2011's first quarter to a range of 3% to 4%, and to estimate that core CPI will come in below 3.5% this year and dip to 3% in 2012. This relatively benign outlook suggests Mexico's central bank won't act to tighten this year—a view that has helped support stability in its currency.
Don't let troubling headlines overwhelm the good
Despite these positive fundamentals, many in the investment world are still wary of Mexico, in large part because of the drug-related violence. The Mexican government reportedly estimates that fallout from all the drug-related killings—more than 34,000 the past four years—shaves 1.2 points of potential GDP growth every year. Just this week, gunmen attacked two U.S. customs agents, wounding one and killing the other—the second murder of a U.S. government employee in Mexico in a year.
As bad as the violence is, however, it is worth noting that the drug trafficking is predominately limited to the northern states; that tourists are still coming—revenue from tourism, a major industry in Mexico, grew 5.3% last year compared with 2009, when the global financial crisis and the H1N1 flu virus kept visitors away; and that the per-capita murder rate in Mexico is about half that of Brazil, where violence gets nowhere near the attention in the U.S. media. This is somewhat of a sore spot in Mexico, where officials feel their country is overshadowed by their South American rival.
Clearly, the Mexico of today is much different than the Mexico of just a decade ago. In its capital city, most cars on the road are relatively new—when portfolio manager Geoffrey Pazzanese visited a dozen years ago, most of the cars were 10- to 20-years-old and of limited variety. He also noticed during last summer's return visit that the streets were clean and manicured, glass towers were everywhere and global brands were being advertised all over. Granted, these are just isolated observations from a single city, but they reflect a view that Mexican officials drove home in discussions and meetings Pazzanese took in during his visit—that Mexico is committed to continued political, economic and societal reforms aimed at improving its competitive position in the world. Sure, the violence detracts from this effort—but it doesn't negate it. Fundamentally speaking, we like Mexico.
Barclays Capital, Emerging Markets Weekly
Developments within emerging markets have been overshadowed by external events.
Evidence continues to mount that the global economic recovery remains intact. The ECB signalled that a rate hike is imminent.
But turmoil in the Middle East and North Africa remains foremost in the minds of EM investors. Oil forwards are pricing a higher oil price in the short term, followed by normalization over the next 3-6 months, at levels consistent with Barclays Capital’s forecasts.
EM assets have been already shown some signs of fatigue; however, investors should be prepared to take advantage of mispriced opportunities, especially for ‘crowded trades’ that have underperformed.
Turmoil in the MENA region continued to drive market sentiment, and price action this past week blurred some more positive news flow, especially from Europe (strong PMI readings) and the US (ISM manufacturing reached new highs, and the ADP survey of employment and the weekly report on unemployment claims signalled meaningful improvement in labor market conditions). A violent and potentially long-lasting standoff in Libya has at least temporarily disturbed oil markets and reinforced fears about prolonged oil supply disruptions from that country and elsewhere. Actual supply disruption has been limited so far. But it occurs within a market context of rather limited excess capacity in the oil complex and amidst anxieties that destabilizing political events could escalate in other large energy producing countries both within the region (particularly, Algeria, Iraq and Saudi Arabia) and without (we take particularly seriously the risk of supply disruptions in Nigeria as that country gears up for elections next month). Given high uncertainty, further political unrest may lead to higher oil price volatility in the coming months). So far, the market seems to be pricing a temporary period of high supply risk, with limited implications for the longer term. The forward price curve has been flattening and inverting in the past few weeks, with the long end moving only marginally higher (Figure 1 shows that WTI, the curve for Brent, is also inverted, with 2012 prices of about $112/barrel). By 2012, the forward curves are comparable to our current forecasts of $105/barrel (Brent) and $106/barrel (WTI).
So far, therefore, political events in the Middle East seem to have validated and brought forward market pricing to Barclays Capital’s longstanding bullish call on oil markets, rather than created an oil-price shock that exceeds what we considered plausible in light of underlying supply-demand dynamics. That said, heightened geopolitical stress seems likely to create mainly upside risks to oil prices in the immediate future, and the abrupt shift in oil prices seems likely to keep questions about the effect of higher oil prices on global growth and inflation at the forefront of investors’ minds during the potentially volatile months to come. But there are downside risks to commodity prices as well. In some of those emerging economies where monetary tightening is most advanced, including China, Brazil, and India, there is evidence that economies are responding to the tighter conditions. It is not hard to imagine an anxiety trade in the not-too-distant future about the effect of the associated deceleration in demand on commodity markets. Such movement could further be compounded if global monetary policy moves to a somewhat tighter stance, as hinted by the ECB’s communication and the anticipated June end to QE2.
The political upheaval in the Middle East seems to have increased uncertainty about oil prices (and associated global economic developments), at least as much as it has increased expected oil prices. The uncertainty is compounded by questions surrounding the policy response to higher oil prices. Could monetary and fiscal authorities in developed countries turn even more dovish as they focus on the negative growth implications of the oil price shock? At least in those emerging economies where inflationary pressures are contained and the spill-over from oil to consumer prices is smaller, that could be the actual policy bias.
What we think
EM assets entered the oil supply shock in a somewhat precarious situation. Retail flows into EM dedicated funds (both fixed income and equity) have been showing signs of fatigue in recent weeks (although institutional flows, according to our read of anecdotal evidence, have remained healthier). Together with inflation fears, crowded positioning and stretched valuations, this situation has accentuated the recent underperformance of EM assets versus their developed markets’ counterparts. ‘Crowded trades’ were particularly penalised. The key question is whether the risk premium already priced into EM assets is sufficiently high. To shed some light on this issue, we analyzed the performance of EM credit, rates and FX against key global drivers. We regressed daily changes of 5y CDS, total return local bond indices and relevant FX crosses against daily changes of VIX, the price of WTI oil, the 10y real UST yield (only in the analysis of CDS and rates) and EUR/USD (only when modelling FX). If EM credit, rates and FX underperformed what our model indicates is fair, we argue that excessive risk premium was priced in. According to this analysis, in the EM FX space, the KRW, COP, TRY and EURPLN have overreacted meaningfully to the recent volatility (the underperformance of the COP probably also reflects the authorities’ extension of the FX intervention program through at least June 17). We think this sets up the KRW, one of our long-standing favourites in EM FX space, for a period of appreciation, now that the technical position has been cleaned up during the recent sell-off.
In the credit space, our model seems to suggest that Argentina, Turkey and MENA CDS have widened more than their betas to global drivers would predict. However, there may be genuine idiosyncratic peculiarities at play. The situation in MENA is obviously unique, with geopolitical risks that are not captured by our simple market analysis. In Turkey, the balance of risks is unclear once we consider the potential effect of oil prices if they remain at current levels. In Argentina, the local news flow has been somewhat disappointing. Negotiations with the Paris Club appear to be moving slowly; hopes for an early normalization of inflation indicators have been dashed; and diplomatic conflicts with the US have reignited anxieties over the increased prospects of a new Cristina Fernandez de Kirchner presidency. Having said this, we remain convinced of the sovereign’s willingness and capacity to pay and find value in short bonds (Boden 15) and front-end CDS.
What we like
As suggested by our analysis, the KRW, not surprisingly, has overreacted to the spike in volatility of the oil price. We recommend taking exposure via an RKO option. This call structure insulates investors against tail risks from any spats with North Korea (or another spike in oil prices), while the RKO feature makes the option more affordable by selling away the upside, which is very unlikely to materialize in this environment, in our view. We expect the BoT to maintain its current pace of tightening. Inflation seems manageable and is being driven by food prices, which appear to have peaked. The market looks increasingly positioned for a more hawkish BoT. We believe the recent steepening in the front end of the curve is somewhat overdone, and we expect the curve to flatten in the coming weeks. Argentina credit, notably a ‘crowded trade’, seems to have overreacted to the recent episode of risk aversion (as noted above). The cash credit curve has inverted, with Boden 15s trading more than 100bp over the USD-Discount. This is an attractive relative value opportunity and we recommend switching into shorter duration Boden 15/Bonar 17 or the short part of the CDS curve. The sovereign's high willingness to pay should be a further catalyst for a normalization of the credit curve (flattening), in our view. Furthermore, the maturity of the Bonar 11 at the end of this month could drive the short-end spread further down, since we expect some investors to extend duration, creating a technical bid on that segment of the curve.
South Africa issued a new USD 750mn 30y bond, extending the existing yield curve out to 2041, as a pre-financing for FY 11/12 (which targeted USD 1bn in external debt issuance). The SOAF 10s30s spread curve is among the steepest in Global EM, and we recommend switching from expensive short-dated bonds (particularly euro-denominated EUR 13s, EUR 16s) into the new issue.
MACRO OUTLOOK: EMERGING ASIA
Diverging growth trajectories
In China underlying growth momentum has moderated, given aggressive policy tightening.
Korea continues to see a resurgence in activity on the turnaround in the global electronics cycle. We expect Bank of Korea to hike its policy rate by 25bp next week.
Next week we get a wave of data from China. We also expect the central banks in Korea, Malaysia and Thailand to hike their policy rates 25bp.
Continental economies – some moderation
Growth momentum appears to be easing in the continental economies. In China, the NBS manufacturing PMI eased in February, in line with seasonal patterns, and driven by cold weather and Chinese New Year holidays. However, looking beyond seasonal distortions, there has been a moderation in underlying momentum since last December, given aggressive policy tightening. Businesses have also been reporting very tight credit conditions, much higher borrowing costs and surging input costs. We think a continued, albeit moderate, slowing of activity in Q1 11 will be healthy and help to ease inflation pressures, but in our view it should still give rise to real Q1 GDP growth of above 9.5%y/y. Going against the regional trend, India’s Q4 10 GDP grew a bit less than expected, at 8.2% y/y. Manufacturing growth remained subdued, reflecting continued pressure on liquidity, rising borrowing costs and margin pressures. So far, GDP growth has been robust, driven by strong domestic demand. Despite persistent high inflation, we feel consumption – the major growth driver and largely non-discretionary in nature – will remain steady, reflecting rising disposable incomes and government spending directed at the rural poor. Investment, on the other hand, is likely to face headwinds, such as tight liquidity, rising borrowing costs, negative headlines about alleged high-level corruption, and global uncertainties.
Indonesia: Softer inflation print, risks that BI may stand pat Indonesia’s February CPI inflation came in below expectations, at 6.8%, following two consecutive months of 7%-plus prints. With core inflation rising to a 13-month high of 4.4% amid elevated inflation expectations, we maintain our call for a 25bp hike to 7%. However, we believe there is an increased risk that the central bank may stand pat, especially given the appreciation of the IDR and dovish comments from BI Senior Deputy Governor Hartadi.
Korea: Robust activity and high inflation – BoK to hike next week
Growth in Korea continues to accelerate. Industrial production expanded for a third consecutive month, rising 4.6% m/m in January. With the inventory/shipment ratio falling to 0.9x and strong external demand, especially for electronics, we expect IP momentum to accelerate further. At the same time, the resurgence in activity continues to feed into demand-side pressures, with core inflation accelerating to an 18-month high of 3.1% in February. Given this backdrop, we maintain our base case that the BoK will hike the policy rate 25bp on 10 March to 3.0% and that the current pace of rate normalisation (a hike every two months) will be maintained, which would take the policy rate to 3.5% by July. The key tail risks to watch in the coming weeks are any deterioration in the situation in the Middle East, which could keep oil prices at higher levels.
Week ahead: China data, central bank meetings in Korea, Malaysia, Thailand
Next week we get a wave of data from China. We forecast the February CPI will rise less than 5% y/y because of slower food inflation, although input costs continue to rise. Fixed asset investment growth is moderating, but remains supported by public housing construction. We also forecast February IP growth will moderate to 24.1% y/y. We expect India’s January industrial production growth to have remained subdued, at 2% y/y, given pressure on liquidity and interest rates. The central banks in Korea, Malaysia and Thailand hold policy meetings. We expect all three to hike their policy rate by 25bp and maintain a front-loading bias, owing to generally strong economic activity and rising core inflation pressures. In Sri Lanka we expect the CBSL to stand pat given the dovish comments from its governor.
MACRO OUTLOOK: EMERGING EUROPE, MIDDLE EAST AND AFRICA
Assessing the impact of oil PMI and IP readings show continuing strength of the EMEA industrial sector.
However, rising energy prices will weigh on the C/As in the non-oil producing areas of EMEA, and they also will put pressure on manufacturer margins.
Also, investment has so far failed to recover, raising long-term concerns about the sustainability of the region’s growth following the global crisis.
EMEA growth improving, with IP taking the lead
EMEA continues to post solid growth performance, implying that export-driven growth remains in place as we approach the end of Q1. February PMI rose on average, increasing in four out of six countries. Manufacturing output and employment contributed the most to the improvements in PMI, while new orders and new export orders posted mixed readings, implying there may be some weakness in the pipeline. Overall, the strong performance is in line with the solid euro area and German PMI, which reached 10-year highs. The Czech outlook remains the brightest, notwithstanding a slight moderation in February, with PMI at 59.8 versus an historical high of 60.5 in January. Notable improvements were recorded in Hungary (to three-year high of 57.0 from 54.7 in January), Russia (to 55.2 from 53.5) and Turkey (to an all-time high of 58.5). Poland PMI was weaker at 53.8 versus 55.6 in January and, along with other indicators, implies deceleration in growth during the first months of 2011.
Industrial production releases next week in Hungary, Czech Republic and Romania will likely extend these positive trends. In South Africa, a slight uptick in PMI was complemented by strong growth in vehicle sales, up by +25.2% in February. This adds confidence to our view that SA industry will leave behind its weak December performance (0.2% y/y vs 4.6% in November) and re-accelerate in January.
Commodity prices fuel C/A deficits and weigh on manufacturing margins
The growth outlook of EMEA faces several challenges. As domestic economies begin to recover and imports rise from their lows, C/A balances are turning weaker. This has been exacerbated by higher commodity prices as food (mainly grains) and energy (oil), have risen. If the unrest in the Middle East is prolonged and spreads to other large oil producers, oil prices could rise further. Higher oil prices have a direct impact on C/A balances and an indirect impact on GDP). As oil exporters, Russia and Kazakhstan stand to gain while other countries stand to loose (Figure 2). Turkey and South Africa trade balance were weaker than expected this week and we expect C/A balances in Israel, Czech to weaken further. Commodity inflation is also causing disparities between input and output prices. The spread between CPI and PPI is putting pressure on producer margins that could affect growth prospects (while of course benefiting extraction sectors, Figure 3). In addition, the disparity is leading to upward pressures on CPI prices. EMEA inflation trends have been mixed. Turkey headline inflation this week continued to decline in line with expectations to 4.2% y/y in February (from 4.9% in January). However, this was helped by strong base effects and PPI inflation is back in double digits (11% y/y in both Jan and Feb 2011). Also, Kazakh inflation was higher. Weak domestic demand may still contain inflation trends in economies such as Hungary and Romania, but we expect higher inflation readings in Czech, Serbia and Egypt next week.
Investments are another factor that could raise concerns. Investment typically lags when an economy comes out of recession (Figure 4). However, thus far, investment has not yet turned, in contrast to the 1997/98 crisis, when investment recovered relatively soon. Average investment-to-GDP ratios in EMEA still exceeded those after the 1997/98 crisis, reflecting their elevated levels before the crisis. While this may signal a ‘new normal’ of lower investment following a long boom in EMEA, a continued weakening of investment would add to concerns about the sustainability of the growth recovery, which remains highly reliant on external demand.
MACRO OUTLOOK: LATIN AMERICA
As the Middle East jitters remain in the spotlight, we think the risk of political contagion in any of the market-relevant economies in LatAm is extremely limited.
The presidential elections in Peru and Argentina this year are another reflection of the consolidation of democracy in the region.
As developments in the Middle East remain in the global spotlight, a salient concern among EM investors has been to identify their potential implications for other emerging markets. In Latin America, we have recently elaborated on the potential economic effects of higher oil prices and, possibly, weaker global demand. However, we have not addressed the potential political spill-over – an issue about which investors seem less at ease than we would have thought. We think the risk of political contagion in any of the market-relevant economies in Latin America is extremely limited. This is the case, in our view, even for those regimes that are perceived as more authoritarian and about which investor queries seem concentrated. The two key distinctive socioeconomic and political conditions highlighted by our EMEA colleagues as facilitating the recent Middle East events have been sky-high youth unemployment and a lack of voice and accountability. Latin America does much better on both accounts, with youth unemployment about 10 points lower on average than in the Middle East and ranking substantially higher in the World Bank Voice and Accountability indicator (Figures 1 and 2.) That the region is generally democratic is also evident in the political news flow, with the completion of three peaceful democratic transitions in Chile, Colombia, and Brazil last year having added credibility to the continent’s economic renaissance. Even Venezuela has just moved to a National Assembly no longer controlled by President Chavez.
There are, of course, exceptions. Cuba, for example, ranks in the fourth percentile of the World Bank’s Voice and Accountability indicator. Also on the fence, Haiti, Nicaragua, Honduras, and Guatemala rank below the 40th percentile. But none of these economies has the gravitas to be destabilizing or elicit more than a shrug from market participants, as the Honduras 2009 coup demonstrates. It is then no coincidence that whatever investor concerns we have perceived since the outburst of revolutionary flares in the Middle East have concentrated on Venezuela in particular – an obviously relevant economy and which is also largely perceived as authoritarian. Interestingly, Venezuela’s self-perception regarding the state of its own democracy differs from the outside perception. According to Latinobarometro’s latest poll, Venezuelans consider their country to be slightly more democratic than what the average Latin American thinks the region is and as democratic as he/she thinks the US is. Moreover, the view of democracy as the best government method is at historically high levels and has increased from what it was during the more unstable years of the Chavez government (2002-04). After the opposition tried and failed to push Chavez out through expedited means in those years, it is now more mature, having recently agreed to hold its own primaries at some point between next November and March. Since the beginning of the year, President Chavez and leaders of the opposition have been singularly focused on the forthcoming election. In this setting, the chances of contagion from the current situation in the Middle East appear low. On the contrary, the country clearly benefits from the increase in oil prices. Moreover, since polls show that the opposition might win, the probability of a democratic transition in Venezuela is increasing. The presidential elections in Peru and Argentina this year are another reflection of the consolidation of democracy in the region. In the case of the former, five weeks before the election, four candidates appear to have a meaningful chance to make it to the second round. Even anti-model candidate Humala exhibits clean democratic credentials, participating in an open and fair campaign. In the case of the latter, President Cristina Kirchner inaugurated Congressional sessions on Tuesday, a period that will be marked by a contested presidential campaign. While she has not yet announced her candidacy, and primary elections have not yet taken place (two of the opposition parties have scheduled them for April), all polls suggest that she enjoys a high probability of being re-elected). It is also important to note that the new administration is unlikely to control Congress.
STRATEGY FOCUS: LATIN AMERICA
Mexico: A sweet growth-inflation mix?
We think the recent combination of growth, inflation, and expectations data is unlikely to disturb Banxico’s contentment with current conditions, despite the riskier global inflation backdrop. We see some value in short-end real rates.
When Banxico’s board convenes this Friday, 4 March, it will face a growth-inflation backdrop that, at first sight, seems enviable. First, activity data should have confirmed the bank’s more upbeat views. Second, inflation, as anticipated by the board, has declined in y/y terms. Finally, inflation expectations, despite the increased global inflation fears, have remained surprisingly unaffected. Underlying these Goldilocks conditions, however, we think that inflation risks have increased, in part because activity has defied our more downbeat outlook and in part because beneath the “benign” headline inflation rate lie trends that hardly seem reassuring. However, with inflation and expectations certainly not pressuring the bank, we do not expect Banxico’s tone to change abruptly or signal any proximate monetary policy move. As our global strategists have pointed out, we think this backdrop should bode well for the performance of linkers, although, admittedly, their liquidity is problematic.
Strong Start to 2011 Activity
The latest round of activity data from Mexico has been quite strong, contradicting our expectation that sequential economic growth would stabilize near Q3’s modest level (approximately 3% q/q saar). Although we still think that some of our caveats about the economic outlook remain valid and preclude a repeat of the strong bounce in 2010, the pickup in activity during Q4, along with a solid start to Q1, leads us to project a stronger full-year reading in 2011 of 3.9% (from 3.3% previously), close to that of the consensus (Banamex: 4.2%). Real GDP expanded 4.6% y/y in Q4. At the margin, it grew 5.1% q/q saar, up from 3.2% in the preceding quarter. The production breakdown showed an evenly strong performance in the secondary (5.3% q/q saar) and tertiary (5.0%) components. Meanwhile, January high-frequency data point to a solid start to Q1. Both export and import flows were robust, expanding 3.7% and 3.3% m/m sa, respectively. This implies acceleration in the underlying trends to 33.2% and 25.0% 3m/3m saar, respectively. Unemployment took a leg down to 5.23% sa, from 5.47% previously. We estimate IP to have expanded 5.6% y/y, consistent with a roughly flat m/m reading but a solid underlying trend of 3.4%. Overall, we now think Q1 may sustain a somewhat firmer pace of expansion than we were expecting (even if probably below that of Q4) and update our estimate to 3.5%, from 3.0% previously.
We view the risks to our revised projections as limited and balanced. After the wave of optimism about US and Mexican growth at the end of last year, solid incoming data have failed to generate meaningful market moves (at least in the FX market), and we continue to see upside potential as limited against a backdrop of already-upbeat expectations. Rather, we emphasize the potential inflation implications of the recent strength in activity data, insofar as it should increase the central bank’s confidence that the output gap is closing quickly. We find it noteworthy that although until Q2 10 the demand-side composition of the recovery had been marked by an outperformance of exports relative to domestic demand (particularly gross fixed capital formation, the production side of the recovery has been much more balanced, with tertiary activities bouncing as strongly as secondary ones. This may appear somewhat counterintuitive; one might think that external demand should spur growth outperformance in the secondary sectors (dominated by manufacturing), while domestic demand growth should cause tertiary sectors (dominated by services) to outperform. We suspect this may reflect the fact that several tertiary sectors, such as transportation and communications, are associated with foreign trade; similarly, some secondary activities (ie, construction and utilities) are probably more sensitive to domestic than external demand. In any case, the bottom line, in our view, is that the production breakdown does not appear to support the notion that there could be significant slack in the tertiary sectors, even if not in the secondary sectors. Indeed, the Q4 report has brought both aggregates close to their respective long-term “trend.”
The Silence of the Sea
On the inflation front, we continue to think that the declining trend is the result of temporary factors and base effects, paving the way for a bounce in y/y inflation as of Q2. Indeed, core inflation m/m readings point to acceleration at the margin, and we think the recent increases in international commodity prices pose some upward risk to our year-end projections of 4.2% and 4.0%, respectively, for headline and core rates.
The pressure from commodity prices has been evident in recent releases, albeit contained. The recent 2w/2w increases in processed food prices have been higher than seasonality would justify, but not significantly so (Figure 3). Meanwhile, energy prices have kept creeping up broadly in line with our expectations. This is consistent with the government’s gasoline pricing policy aimed at closing the gap between domestic and international prices in a gradual manner. We do not expect the recent spike in oil prices to lead it to accelerate the pace of gasoline price hikes significantly, despite the recent international spike, but we think the latter entails pipeline pressures that would warrant the continuation of the price adjustment in 2012 if the international shock does not unwind (Figure 4).
Banxico’s next steps
We do not expect this growth-inflation backdrop to shake Banxico from its relatively unworried tone. It was already expecting ongoing dynamism on the activity front, and for now, inflation data have not contradicted its projections. Moreover, inflation expectations in Mexico have been remarkably contained, even as inflation has re-established itself as a dominant concern for EM investors globally Finally, the signals from the factor markets that Banxico follows to assess overheating pressures (eg, consumer credit, wage growth) have been improving, but hardly seem ebullient (Figure 6). With the bank failing to signal a change in stance, we think the short end of the curve should remain relatively anchored. Against this backdrop, and given our inflation views, we think inflation-linked rates should do well, as our global strategists have emphasized, although liquidity could be problematic
STRATEGY FOCUS: HUNGARY
Still too many questions to turn positive
The Hungarian government has unveiled its eagerly awaited fiscal reform plan. While announced headline saving numbers are substantial, the plans revealed this week still leave many details open. Uncertainty relates to the realism of the assumed expenditure savings for the various measures as well as the implementation (mainly to occur only in 2012). Therefore, markets’ cautious initial reaction following the announcement on Tuesday seems reasonable, in our view, especially when considering the substantial rebound of Hungarian assets in the run-up to the fiscal announcement. We maintain our defensive stance on Hungary cash credit and reiterate our recommendation to pay 5y IRS. The Hungarian government unveiled its eagerly awaited fiscal reform plan. The plan assumes that the fiscal balance will improve by HUF556bn in 2012 and HUF902bn in both 2013 and 2014. This translates to 1.8% of GDP in 2012 and 2.8% and 2.6% of GDP in 2013 and 2014, respectively, depending on the precise growth and inflation forecasts. As reported by the press ahead of the announcement, the overall headline numbers of HUF900bn for 2013 and 2014 were slightly higher than the initially communicated HUF600- 800bn. As the government indicated in its press conference, roughly three-quarters of the targeted savings for 2013-14 seem based on 'structural reforms' on the expenditure side. The additional measures on the revenue side, particularly a delay in CIT cuts and an extension of the extraordinary bank tax at full rates, bring the total numbers to the headline HUF900bn in 2013 and 2014. Indeed, thus far these revenue measures can also be seen as the most concrete of all the measures announced. The delay in the corporate tax cut seems like a reasonable decision in light of the fiscal pressures – and may also be a decision welcomed by the EU (think about Germany's and France's attitude towards Ireland's corporate tax and generally towards the idea of competition among EU states via low CIT rates). However, leaving the already comparatively high banking tax fully in place for another year will not help the recovery of lending activity. This could not only weigh on growth, but also potentially accelerate foreign banks’ reduction of exposure to Hungary (ie, reduced foreign debt rollover).
Clearly, the fact that the Hungarian government is actively addressing the longer-term fiscal issues is a positive and has also greatly alleviated the risk of a downgrade to non-investment grade status by the rating agencies in the short term. However, while headline saving numbers are substantial and seem broadly in line with the (enhanced) targets reported previously, the plans revealed this week still leave many details open regarding the implementation of announced measures, particularly in light of the potentially adverse public opinion ahead of the 2014 elections. Measures such as cutting unemployment benefits, cutting drug and public transport subsidies and reducing education spending are achievable in principle. However, no details about the specific measures to reach the saving targets were provided and it remains difficult to assess how easily the purported savings can be made. Moreover, some of the underlying assumptions also lack clarity, in our view, and some of the government’s targets seem quite optimistic: eg, the goal to achieve 4-6% real GDP growth and create an additional 300,000 jobs by 2014. Also, according to reports from the press conference, the government is apparently counting on more favourable funding costs over the next years, despite Hungary's high debt redemptions and issuance needs, both on the local and external debt side (Figure 3), and despite the global environment of higher inflation and rising core rates. This leaves a relatively large degree of uncertainty about the future debt dynamics.
In a simple baseline scenario, where we see an interest rate-growth differential of 100bp (7% nominal interest rate and 6% nominal GDP growth) and leave the exchange rate unchanged, we find that in a scenario where fiscal policies had been left unchanged at the decisions taken up to end-2010, the debt-to-GDP ratio would start to rise after 2013 and reach 85% by the end of the decade. The immediate drop in the debt ratio in 2011 is driven mainly by an assumed 5% of GDP reduction in debt as a result of the re-nationalised private pension assets. If the government was to implement the measures in full (and the savings are as announced), we see debt to GDP reaching the Maastricht level of 60% by the end of the decade. Thus, in reality, investors could well face a scenario somewhere in between, ie, one where after its one-off drop in 2011, the debt ratio remains at elevated levels without a clear downward trend.
Cautious on Hungary cash credit, recommend paying 5y IRS
The market has eagerly been awaiting the announcement of the planned reform measures and, judging by the outperformance of Hungarian assets YTD, investors seemed to have given Hungary the benefit of the doubt. There are new positives in the announcement but the remaining question marks over implementation and realism of the fiscal package have likely contributed to the rather lukewarm initial market reaction on Tuesday. In the credit space, we maintain our Underweight stance on Hungary cash credit and reiterate our recommendation to sell the 10y basis (Hungary USD 20s, 10y CDS). Given valuations (especially on USD cash bonds), the heavy supply outlook and the remaining uncertainties over the longer-term fiscal outlook, we expect Hungary cash credit to underperform from current levels, both versus other regional credits and against CDS. In Hungary local markets, the local yield curve is still pricing in only 20bp of rate hikes up to the 2y sector and we think more rate hikes should be priced in. Over the past few months, foreign holdings of HUGBs have increased rapidly, partially on positive expectations regarding the fiscal reform programme announcement, and partially driven by the stabilisation in sentiment towards peripheral Europe. The latest data suggest that HUGBs are no longer underweighted by international investors and we think that the rally is likely to lose its momentum. We reiterate our recommendation to pay 5y IRS.
STRATEGY FOCUS: INDIA
India’s FY11-12 budget deficit forecast of 4.6% of GDP is markedly lower than market expectations. We see significant risks of an upside surprise to the number. Despite the lower forecast deficit, net supply of SLR securities remains high relative to our demand estimates, and we look for G-sec yields to move towards 8.40% in H1 FY11-12. Government sees headline fiscal deficit at 4.6%
India’s government is projecting a budget deficit of 4.6% of GDP in FY11-12, markedly lower than market expectations (BarCap: 5.3%). While the government's revenue (both tax and nontax) forecasts are very close to our estimates, the expenditure figures are much lower than our expectations. We estimate that the 4.6% headline fiscal deficit target implies net and gross market borrowing of INR3.43trn and INR4.17trn, respectively. The FY11-12 net borrowing requirement is also at a level that is strikingly similar to that seen in FY10-11. We had expected the government's fiscal deficit estimate for FY11-12 to be based on aggressive assumptions and, accordingly, had forecasted the fiscal deficit at 4.8-5.0%. We see a risk that the government’s expenditure projections could be the source of upside surprises, leading to a higher fiscal deficit. This would also imply that the government’s market borrowing needs would be raised. But we would only expect such revisions to be made in H2 FY11-12. The government’s expenditure estimate looks remarkably low. Overall expenditure for FY11-12 is budgeted to increase a mere 3.4%, markedly lower than the CAGR of 15% over 2000-10. Within this, assumptions for non-plan expenditure are also somewhat surprising, with a budgeted reduction of around INR54bn from FY10-11 levels. Again in contrast with average long-term growth in the mid-teens. Likewise, non-plan capital expenditure is budgeted to shrink ~INR120bn. Growth in plan expenditure is budgeted at 12% for the fiscal year. Also, petroleum subsidies (including bonds) cost around INR800bn in FY09, a period when crude prices (Dubai) averaged USD82/bbl. We believe the expenditure estimates are based on aggressive assumptions and are likely to be revised higher.
Government bonds – near term stable, but supply pressures acute
We expect 10y G-sec yields to be stable over the next month and a half. The lower budget deficit headline number for FY11-12, coupled with a) limited supply until April and b) an increase in SLR requirements due to year-end deposit base increases will act as near-term positives for the bond market. We believe benchmark 10y G-sec yields (ie, 7.80% GoI 2020s) could soften towards 7.90% in the next month. However, we expect G-sec yields to move higher in FY11-12 on supply pressures. Even with the lower fiscal deficit target, we expect net supply of INR4.83trn of SLR securities in FY11- 12, including INR3.43trn of G-secs, INR1.25trn of state development loans and INR150bn of Treasury bills. Of this total, we expect commercial banks to absorb around INR2.50trn (52% of net supply). However, we estimate that demand from all other sources will only be around INR1.90trn. This implies a net demand shortfall of around INR400bn for SLR paper in FY11-12, even if the central government sticks to its current borrowing target. The pressure on rates will be more acute in H1 FY11-12. Typically, the government issues around 65% of its bonds in the first half of a fiscal year, while only around 40% of banks' full-year deposit growth tends to happen in the first half. Assuming full-year deposit growth of 16.5% and an excess SLR buffer of 3%, we estimate the banks will absorb around INR2.50trn of the net supply, of which only around INR1.0trn will be in H1. This compares with expected net supply of around INR2.70trn in this period. In fact, putting all of the sources together, we think there will be an excess supply of over INR550bn in H1 FY11-12. Even if we assume the RBI conducts some OMOs and buys back around INR200bn – the amount we estimate is required to bring the system liquidity deficit to within 1% of NDTL – there would remain an excess supply of over INR350bn in H1. We expect the RBI to hike policy rates by a further 75bp in 2011, and we think that banks will be reluctant to add duration in a rate-hiking environment. Higher oil prices and persistent high inflation will act as further negatives and we expect to see 10y G-sec yields moving towards 8.40% in H1 FY11-12.
The situation in H2 FY11-12 will depend largely on whether the government sticks to its budgeted issuance. If the government increases its borrowings (a high probability in our view, barring one-off revenue gains), we would expect bond yields to continue to move higher, despite the RBI nearing a pause. However, if the government sticks to its borrowing programme, we think demand-supply dynamics will likely stabilise bond yields as SLR demand from banks will likely increase. However, the mismatch will remain on a full-year basis, and we find it difficult to see yields heading lower in 2011 – unless there are significant changes in the current liquidity, inflation and growth dynamics.
Risks to this view:
Support from the RBI in the form of open market operations (OMOs) is the key risk to our view. Our projections indicate that banking system liquidity will remain in deficit well into FY11-12. If the banking system deficit increases significantly beyond the comfort zone of the RBI (1% of NDTL), we think the central bank would inject liquidity via OMOs, offering potential support to the bond market. We estimate that the RBI will do around INR500bn of OMOs in FY11-12. However, it is worth noting that the impact of such operations on yields will be very much depend on the timing of such activities.
Given the government’s high cash balance, it may issue fewer G-secs in H1 (we estimate that around 65% of total issuance will be in H1, as per historical patterns). Also, the government can rely on the ways and means advances (WMA) provided by the RBI. We expect flattening in the OIS curve to continue in the near term As mentioned above, near-term relief in bond yields should mean relief for 5y OIS as well. At the same time, we see other factors keeping the front-end of the OIS curve elevated: a) structurally tight liquidity due to currency demand; b) seasonal tight liquidity due to tax outflows in March; c) increased T-bill issuance in FY11-12; and d) rate hike expectations. We continue to expect the OIS curve to flatten and look for 2x5 OIS to move towards 40bp. (Current 50bp, Entry, 58bp, Stop-Loss 68bp).
Risks to this flattening view largely arise from lower growth outcomes. If high inflation and elevated commodity prices, coupled with tight liquidity conditions, affect future growth, the central bank might follow an easier path on monetary tightness than is currently priced in. However, for now, we think inflation concerns far outweigh growth concerns and attach a low probability to this scenario.
Michael Jabara, Stephen Minar, and David Perlaman, Morgan Stanley Smith Barney
ETF Tracking Error Declined Materially in 2010
Tracking error declined across all market segments and virtually all providers for US-listed ETFs in 2010 and averaged 74 bps. We define tracking error as the difference in total return between an ETF’s net asset value (NAV) and its underlying index. In our view, the most common sources of tracking error include fees and expenses, portfolio optimization, and index changes. Additionally, compliance with IRS/SEC diversification requirements can lead to extreme tracking error for select ETFs as they may be forced into material weighting and holding deviations from their stated benchmarks
We found a narrower range and magnitude of tracking error in 2010 versus 2009. In 2010, the range of tracking error fell 1,125 bps to 584 bps and 67.7% of ETFs exhibited lower tracking error year-on-year with an average decline of 88 bps. In addition, the percentage of ETFs with tracking error greater than 100 bps decreased from 37.1% in 2009 to 22.0% in 2010. Conversely, the percentage of ETFs with tracking error less than or equal to 25 bps increased from 22.6% to 26.4%.
Fund expenses as a percentage of weighted average tracking error increased from 31.7% to 62.8% in 2010. In our view, this is a constructive development for ETFs. A higher ratio of expenses to tracking error corresponds to more effective tracking of index performance. Yearon- year, the weighted average expense ratio remained level at 32 bps while weighted average tracking error decreased from 101 bps to 51 bps (-49.5%). In addition, the percentage of ETFs with tracking error below expenses rose from 27.0% to 32.0%.
International ETFs accounted for 38.0% and 57.9% of average and weighted average tracking error, respectively, in 2010. Conversely, International ETFs comprised just 25.5% of ETFs included in the study.
Tracking error is a concern to ETF investors. ETFs are designed with the objective of providing access to and replicating the performance of specific indices. One of the primary goals of ETF portfolio managers is to minimize tracking error, which we define as the difference in total return between an ETF’s net asset value (NAV) and its underlying index. Therefore, tracking error can indicate to us how well managers have met their objectives. We note that our approach to evaluating tracking error may differ from other calculation methods.
Factors such as fees and expenses, diversification requirements, portfolio optimization (holding a representative sample as opposed to a full replication), index turnover, and dividend reinvestment policy may cause an ETF’s NAV performance to deviate from its underlying index. Additionally, for ETFs providing exposure to less liquid markets, such as emerging markets and fixed income, the liquidity and pricing of underlying securities can have a meaningful impact on an ETF’s ability to track index performance.
We analyzed 2010 tracking error for US-listed ETFs. Our analysis is based on the annual total return (including dividends) of ETF NAVs versus underlying benchmark indices. Exhibit 11 contains 2010 tracking error data for each US-listed ETF with an inception date prior to January 1, 2010. Excluded from the study are active ETFs, ETFs that track a static basket such as HOLDRs, physical commodity, and currency ETFs that do not track indices, as well as leveraged, inverse, and leveraged inverse ETFs that have daily return targets.
Observations & Analysis
For 2010, average and weighted average tracking error for all US-listed ETFs was 74 and 51 basis points (bps), respectively. This reflected a material decrease of 51 bps in average tracking error as well as a meaningful decrease of 50 bps in the weighted average tracking error from 2009 levels of 125 and 101 bps, respectively. Exhibit 1 displays the tracking error for each major market segment.
Within US Equity, tracking error decreased to an average of 57 bps in 2010 from 84 bps in 2009. We note that all US Equity market segments saw decreases in their averages. For example, in 2010 US Sector / Industry ETFs exhibited average tracking error of 67 bps, down from 104 bps in 2009. This decline is largely attributable to a drop in tracking error for 63% of the US Sector / Industry ETFs included in last years study; however, we note that one fund reduced tracking error by 631 bps year-on-year and two funds exhibiting 546 bps and 1,709 bps of tracking error in 2009 were removed from this study due to changes in their underlying index. On a weighted average basis, tracking error declined from 27 bps to 24 bps in 2010 for US Equity ETFs. Interestingly, weighted average tracking error increased for four of the six US Equity market segments, while the median tracking error fell for all US Equity market segments. After leading the pack in terms of low average tracking error over the past three years, US Style ETFs dropped to number two among all ETF segments. Average tracking error for the group remained strong at 42 bps versus 54 bps in 2009, and is now just 10 bps above the group’s average expense ratio. In conjunction, US Major Market ETFs posted a 16 bps drop in average tracking error to 47 bps, which brings the group closer to its annual average of 32 bps over the past nine years. Additionally, US Major Market ETFs exhibited the lowest weighted average tracking error of all market segments in 2010 at 19 bps. Notably, the US Asset Allocation segment exhibited the largest year-on-year decline in both average tracking error and weighted average tracking error, declining from 84 to 29 bps and from 125 to 29 bps, respectively.
In terms of average tracking error, US Asset Allocation ETFs now hold the top ranking and represent the only segment where tracking error is lower than expenses. ETFs within the US Custom category exhibited the broadest decline in tracking error (81% of ETFs), while weighted average tracking error for the group actually increased by 4 bps. We attribute this to a combination of a 3 bps increase in weighted average expenses for US Custom ETFs year-onyear and a 5 bps increase in tracking error for the Powershares Water Resources Portfolio (PHO), which accounts for roughly 37% of US Custom assets. Driven primarily by an increase in tracking error of 49 bps and 19 bps for the iShares DJ Select Dividend ETF (DVY) and the SPDR S&P Dividend ETF (SDY), respectively, weighted average tracking error for the US Dividend Income segment rose to 40 bps, the largest absolute increase of all segments. DVY and SDY, combined with the Vanguard Dividend Appreciation ETF (VIG), account for 87% of US Dividend ETF assets. We note that the 40 bps year-on-year decline in average tracking error to 60 bps was mainly due to the closure of one ETF and the reclassification of another, both of which exhibited significant tracking error in 2009. For the second consecutive year,
ETFs based on international indices exhibited the largest absolute level of tracking error in our study. Despite exhibiting the highest average and weighted average tracking error, International ETFs improved materially from 2009 levels. Within the International segment, 93 ETFs, or 72.7%, exhibited lower tracking error compared to their 2009 level, with an average decrease of 152 bps. Additionally, the percentage of ETFs with tracking error below expenses improved from 20.7% to 35.1%. As a result, the average tracking error for the group improved by 84 bps to 110 bps and the group’s weighted average tracking error declined by 128 bps to 104 bps, the largest absolute year-on-year decline of all ETF market segments. We note that a significant factor affecting weighted average tracking error for International ETFs was the 375 bps yearon- year decline in tracking error for the iShares MSCI Emerging Markets Index Fund (EEM), which accounts for over 19% of International ETF assets. In 2010, EEM accounted for roughly 57 bps of the group’s 104 bps weighted average tracking error, which compares to EEM’s contribution of roughly 141 bps of the 232 bps weighted average in 2009. In March 2010, EEM transitioned to a three-basket creation / redemption structure, which allows the fund to more fully replicate the MSCI Emerging Markets Index. During 2010, EEM increased its exposure to local securities by 39% to a total of 68%, and the number of securities held increased by 300. We note that in 4Q10, EEM exhibited just 2 bps of tracking error. International ETFs accounted for 18 of the 28 ETFs with 2010 tracking error greater than or equal to 200 bps. Furthermore, while comprising just 25.8% of ETFs included, International ETFs accounted for 38.0% and 57.9% of average and weighted average tracking error in 2010. Noteworthy, in our view, is the relatively large tracking error exhibited by some ETFs tracking emerging market (EM) indices. For example, 13 of the 18 International ETFs mentioned above track EM indices and averaged 351 bps of tracking error in 2010. We attribute this primarily to optimization techniques, which help to increase liquidity and reduce trading costs and taxes; however, this may increase tracking error. For example, the SPDR S&P Emerging Markets Small Cap ETF (EWX), which currently holds 738 of the 1,667 securities in its benchmark, had negative tracking error of 450 bps in 2010. This was largely attributable to the fund’s relative underweight in smaller holdings, and to a lesser extent, EWX’s country underweights, both of which detracted from performance.
Global ETF tracking error also decreased substantially in 2010. Global ETFs averaged 76 bps of tracking error in 2010 versus 158 bps in 2009 as 76.5% of Global ETFs exhibited lower tracking error year-on-year (average decline of 120 bps). In addition, 54.1% of Global ETFs exhibited tracking error less than fund expenses, compared to 31.5% in 2009. Last, the number of ETFs with tracking error above 100 bps decreased from 40.7% to 26.2% in 2010 while the number of ETFs with 50 bps or less of tracking error increased from 25.9% to 50.8%. We note that the results were slightly skewed by PowerShares Global Listed Private Equity Portolio (PSP), whose NAV return lagged its index by 510 bps (versus 1,368 bps in 2009) due to special tax treatment of private equity companies held in its portfolio1.
ETFs tracking fixed income indices improved significantly in 2010. After ranking 3rd to last in 2009, the Fixed Income segment reduced tracking error by 86 bps on an average and weighted average basis, ranking 4th among ETF segments in 2010. Additionally, the percentage of ETFs with 25 bps or less of tracking error increased to 45% and the percentage of ETFs with 100 bps or more of tracking error decreased to 16% from 30% in 2009. We note that average expenses for the group remain the lowest of all ETF segments at 25 bps. In our view, the primary reason for an 85 bps reduction in weighted average tracking error in 2010 was due to five ETFs comprising 32.5% of Fixed Income ETF assets. In 2009, these five ETFs averaged 480 bps of tracking error and accounted for roughly 74.4% of the weighted average, whereas in 2010, they reduced tracking to an average of 63 bps, thus accounting for only 43.6% of the 39 bps weighted average.
From 2003 to 2006, Fixed Income ETFs ranked #1 among all ETF segments in terms of tracking error (we note that, excluding a preferred ETF, average tracking error in 2007 would have been 13 bps, the lowest of any market segment). During that period, the Fixed Income segment primarily consisted of ETFs tracking Treasury indices. Since then, the Fixed Income ETF market has vastly expanded and in our 2010 study includes 83 ETFs investing in high yield credit, preferred securities, international bonds and municipal bonds, all of which tend to exhibit above average tracking error relative to the group. For example, the seven international, two high yield, 19 municipal bond (national and single state) and four preferred ETFs exhibited 73 bps, 81 bps, 87 bps and 138 bps of tracking error, respectively, in 2010. This compares with just 18 bps for ETFs investing in Treasury securities.
In 2010, average and weighted average tracking error declined by 27 bps and 40 bps, respectively, for Commodity ETFs. The drop was primarily related to the Powershares DB Commodity Index Tracking Fund (DBC), which accounts for roughly 30% of segment assets and whose tracking error fell from 127 bps in 2009 (accounting for roughly 34% of the weighted average) to 9 bps in 2010, which was the lowest absolute tracking error among Commodity ETFs. We also found that a combination of lower asset concentration and mixed tracking error performance among the remaining 15 ETFs led to more muted results for the group.
We found a narrower range and magnitude of tracking error in 2010 versus 2009. In 2010, the range of tracking error dropped 1,125 bps to 584 bps. Moreover, we observed less cases of high tracking error in 2010 than in 2009. In 2010, 4.6% (27 / 592) of the ETFs included in this study exhibited tracking error that equaled or exceeded 300 bps. This compares with 15.6% (88 / 563) of the ETFs we analyzed in 2009. As a result of lower absolute and relative tracking error in 2010, 32% of ETFs exhibited tracking error less than or equal to their respective expense ratios 1 To compared to 27% in 2009. Interestingly, we note that despite above average tracking error levels in 2009, the percentage of ETFs exhibiting tracking error of 10 bps or less increased by just 0.8% in 2010. Exhibits 4 and 5 list the ETFs with the highest and lowest absolute tracking error in 2010.
Our study indicates that average tracking error was positively impacted by a broad improvement among ETFs. Exhibit 2 lists ETFs with the largest absolute decrease in tracking error from 2009 to 2010. On average, the 24 ETFs listed exhibited a 553 bps drop in tracking error. In addition, we note that roughly 68% of ETFs experienced a decline in tracking error year-on-year, with an average decline of 95 bps. This compares to an average increase of only 29 bps for the 32% of ETFs whose tracking error increased in 2010.
Robert Lee, Larry Hedden and Jacob Troutman, Keefe, Bruyette and Woods
Asset Management Distribution Survey: Taking More Risk, but in a Measured Way
We conducted a survey of distributors of retail asset management products over the months of December 2010, January 2011, and February 2011 in order to identify distribution and demand trends for mutual funds and other retail investment products.
For our survey group, we selected more than 600 “gatekeepers” for investment product sales at various distributors. The gatekeeper is often the key contact between the asset management company and the gatekeeper’s respective distribution system and a total of 27 responded to our survey.
That an aging population is likely to accelerate demand for investments with an income and income growth orientation is supported by the responses we’ve received over time. Demand for products with these characteristics predates the bear market of 2008 and in our view is further proof that it is being driven by secular demand.
While increased demand for income-oriented securities and investment products should generally be good for fixed income products and fixed income managers (and presumably annuity companies), we believe there is, and will be, growing demand for investment products and solutions that emphasize not only consistent income generation, but also the growth of income and dividends in order to meet baby boomers’ income growth needs in retirement. These investment solutions will necessarily include equity and other asset classes, possibly even absolute return-oriented investment strategies, and some may have the specific goal of creating a growing lifetime income stream.
Respondents to our survey seem to agree with this view. An overwhelming 74% of respondents to our survey expect demand for investment products that emphasize growth in income to increase over time, consistent with the results from prior surveys. We believe most investors’ investment experience over the past decade (which is to say disappointing), coupled with evolving secular needs, will drive this trend over time, a trend that we think is consistent with the growing demand for asset-allocation type products). Both of these trends we believe contribute to the erosion of style box investing.
Over the past several years a variety of asset managers, including Fidelity, Pimco, Schwab, and Vanguard, among others, have brought to market products with the specific goal of generating consistent income and/or income growth by managing an asset allocation and annual distribution. In hindsight, the timing of some of these new product launches could not have been worse coming as they did just prior to a steep decline in asset values in late 2008 and early 2009.
Nevertheless, we believe development of retirement income-generating products and solutions is the next product frontier in the retail asset management industry and is evidence of the industry’s migration towards a solutions- and outcome-oriented business. In our view, asset managers and distributors that can successfully build solutions to meet the need for generating consistent retirement income and income growth, particularly if a way can be found to do so without the cost and complexity of many annuity products, will be at a long-term competitive advantage relative to peers.
The Growing Importance of Retirement Income Highlights That Product Construction Can Be a Core Competency
The sharp drop in asset values in 2008, weak investment returns over the past decade, an uncertain (but improving) economic outlook, and a baby boom generation staring retirement in the face, have focused many investors’ attention on the importance and need for generating predictable retirement income. In response, many asset managers are examining ways to incorporate retirement income strategies into their products, particularly as it relates to products destined for 401(k) and other retirement assets.
Some managers, such as AllianceBernstein and BlackRock, have rolled out 401(k) products that incorporate separate insurance riders or annuity products in order to provide, for a separate fee, a way for investors to purchase some form of retirement income or guaranteed minimum income benefits. While some of these are not necessarily annuity contracts per se, the idea is to essentially compete against annuities by offering some level of protection (underwritten by an insurance company) with a modest incremental cost.
While conceptually these insurance products or riders fulfill a key need, we think it remains an open question as to how receptive investors and plan sponsors will be to these innovations given their incremental costs and added complexity. We think it’s worth noting that there is likely not enough insurance capacity in the life insurance industry to underwrite the retirement income needs of the baby boom generation, so non-insurance-based solutions will need to be found.
Over the Years Demand for Closed-End Funds Has Exemplified the Demand for Income
The drive for investment solutions designed to meet a growing demand for income has, over the years, been exemplified by what was, until 2008 at least, robust demand for closed-end funds (Exhibit 22). Closed-end funds tend to have an income orientation and are generally sold on the basis of current yield. To a large extent, many of the investment strategies used in equity closed-end funds are essentially designed to convert total return into an ongoing income stream through the use of options, leverage, and/or derivative strategies, among other things.
Since most income-oriented closed-end funds are focused on maximizing monthly dividends, they tend not to have a particular income growth orientation. It is also difficult to have an income growth orientation in a closed-end fund (or even a traditional mutual fund for that matter) since as investment companies they are generally required to distribute virtually all of their income and realized capital gains. Because realized gains (and income) from most closed-end funds are not retained by the fund to be reinvested, unless the fund is willing to pay the taxes due, it becomes difficult over time for a closed-end fund or even a traditional fund to generate a growing stream of income since there is a limit to the capital that can be retained to grow the asset base. Even if distributions are reinvested in a closed-end fund, it would still be difficult to retain distributed capital in a closed-end fund as dividend reinvestment is often done through the purchase of existing shares in the open market, often at discounts to net asset value (NAV), meaning distributions do not flow back into the asset base of the fund. In this sense traditional mutual funds could be superior to closed-end funds.
While closed-end fund issuance has rebounded modestly over the past year or two, the persistent discounts to NAV at which many closed-end funds trade coupled with the debacle in the auction rate preferred market stemming from the credit crunch in 2007 and 2008, suggest to us that new closed-end issuance is likely to remain muted for an extended period of time.
The demand for income growth is also seen in what distributors expect will be increased demand for annuity products (Exhibit 23). Many respondents to an earlier survey expected demand for annuities to sap demand from mutual funds, which is another reason why many fund companies are trying to develop competing income-oriented products. That said, some asset managers stand to benefit from increased demand for annuity products to the extent they sub-advise investment options for annuity and other insurance products.
Demand for Structured Asset Allocation Products Should Continue to Grow
One of the fastest-growing product segments in the mutual fund industry has been asset allocation and target-date (or life-cycle) funds). While there can be vast differences in structure and some intended use, a common theme is that an asset allocation is being managed with the intention of controlling downside risk overtime.
Most target-date funds (and some asset allocation products) are structured as fund-of-funds products, where the assets are allocated among the manager’s existing mutual funds. A new generation of target-date products has open-architecture structures that can incorporate funds and/or commingled accounts from non-affiliated asset managers. In these asset allocation or target-date products, a manager may provide both the asset allocation overlay service (in a target date product it’s called managing the “glide path”) as well as, hopefully, the underlying portfolio management of the various asset sleeves.
In a non-target date product, the manager may follow an “unconstrained” strategy and have the freedom to invest the fund’s assets in a wide variety of asset classes and individual securities, as they deem appropriate. Flows to asset allocation products with these characteristics have generally been very strong and include Waddell & Reed’s Ivy Asset Strategy Fund and BlackRock’s Global Asset Allocation Fund, among others.
From our perspective, the growth of these types of products is yet another example of the industry’s migration to a solutions-based business, and this migration is being driven in part by distributors’ and investors’ desire for simpler, turnkey investment solutions, as well as investment solutions designed to better control risk and limit downside. This is particularly prevalent in the 401(k) market where many plan participants lack the knowledge, time, or inclination needed to make proper investment decisions Accordingly, considering the plethora of investment products that are available, being able to pick one product that automatically rebalances with an investment horizon that corresponds with a specific investment goal greatly simplifies the complexity and stress of investing for many less experienced or less inclined investors.
These types of products, particularly “life cycle” or target-date products are ideally suited for a 401(k) and other retirement plans. Over time, we expect demand for these products to continue to grow as more plans include target-date products in their investment line-up and more plan participants (and sponsors) become attracted to their ease of use. As use of these products continues to expand, it should be supportive of some continued level of flows into fixed income strategies, regardless of the level of interest rates.
While most target-date funds are single manager oriented and basically allocate assets among a given manager’s various fixed income, equity, and other offerings, a newer generation of target-date products have an open-architecture structure that allows the plan sponsor to choose strategies from a variety of unaffiliated asset managers, including separate manager for the glide-path. We also expect that asset allocation strategies will expand beyond simple traditional stock and bond allocations, to include possibly other asset classes and non-traditional absolute return strategies, among other things.
However, we expect open-architecture target-date products will mainly appeal to the largest plan sponsors who can best use skills and leverage resources developed in managing their defined benefit plans. This can create some opportunity for managers without the 401(k) footprint of a T. Rowe Price, for example, to capture a share of these retirement assets in an investment-only capacity. It also means that some fund managers may be hired just to manage a glide path (the asset allocation) even if they don’t manage some or all of the underlying strategies. AllianceBernstein has been pursuing these types of mandates as a way to build its defined contribution business.
Retail Managed Accounts: Expectations Moderate
Over the years many distributors focused on shifting their business model toward a fee-based (as opposed to transaction-based) revenue model. As a result, managed accounts (also called separately managed accounts or SMAs), which come in different forms, over the past decade or so we believe captured a growing share of the retail investment business.
Based on our most recent survey, the expected growth of managed accounts is consistent with more recent surveys and suggests to us that expected growth is moderating. Overall sales expectations remain positive, with 27% of respondents expecting greater than 5% sales growth over time and another 31% expecting modest growth. However, in a change from the past, sales expectations for managed accounts are generally lagging that of mutual funds, at least with regard to the respondents to our survey. On average it appears that respondents expect annual growth in the low-single-digit range. We note that to some degree the results of our survey could be skewed by the large proportion of respondents from smaller, independent broker dealer and advisory firms, whereas large, national broker dealers tend to be greater proponents of managed accounts.
Anecdotally, we believe that flows into managed accounts over the past several years have been negatively impacted by the same factors that pressured equity mutual fund sales, namely a reduced appetite for risk, particularly since managed accounts tend to be equity oriented. Also, as financial advisors changed firms at an accelerated rate over the past several years, there may have been greater churn among managed accounts relative to mutual funds as managed accounts are generally less portable from broker-dealer to broker-dealer compared to mutual funds. In addition, because of their structure and relatively low minimum account size for a separate account, they tend to be more equity oriented in nature with fixed income strategies accounting for a more limited proportion of separately managed account assets. It may also be harder to accommodate risk adjusted strategies or multi-asset class solutions in a managed account.
Another factor that could be impacting the managed account business is the growth of Unified Managed Accounts (UMAs). Essentially a UMA is a fee-based account that can hold a wide variety of investment products and strategies, including individual stocks and bonds, mutual funds, as well as SMAs. As UMA accounts have taken hold, assets they may have moved to a managed account in the past may now be flowing to other investment vehicles within a UMA.
Regardless of its form, we believe many financial advisors will continue to focus on growing their fee-based accounts, and as the broker-dealer industry migrates towards a fiduciary standard of care as part of the new Frank-Dodd Financial reform bill, we think this will only serve to increase demand for fee-based accounts.