Elections

Special Report

Dear Client, I am on the road visiting clients in Toronto, Chicago, and Wisconsin this week, and as such there will be no regular Weekly Report. Instead, we are sending you a Special Report written by my colleague Marko Papic, Chief Strategist of BCA's Geopolitical Strategy service. In this report, Marko argues that Hillary Clinton has not yet sealed the election, despite her high odds of winning. I hope you will find this report both interesting and informative. Best regards, Peter Berezin, Senior Vice President Global Investment Strategy Highlights Clinton has a 65.5% chance of winning the presidency. A Trump win requires a surprise - such as in voter turnout. Still, we doubt Trump can punch more than 3% above his polling. Regardless of the outcome, multinational corporate profits will suffer. Go long the USD. Feature With the conclusion of the final presidential debate on October 19, the U.S. election is now in its final inning. Donald Trump's chances of mounting a comeback are slipping away (Chart 1). Could there be a Brexit-like surprise for the markets on November 8? And what are the investment implications of this year's unprecedented election? How Trump Can Still Win... Paddy Power, one of the world's biggest bookies, has begun to pay out bets to people who had wagered on Secretary Hillary Clinton winning the election. Meanwhile, according to Nate Silver, America's statistical Geek-in-Chief, Donald Trump has a meager 13.7% chance of winning the election.1 While our own model gives Clinton a 65.5% chance of winning, we have not forgotten Yogi Berra's wisdom: "It ain't over till it's over." There are three reasons why we would have held onto the pay-outs if we ran Paddy Power: Turnout assumptions could be wrong: Silver's quant model - and ours - is based on the assumption that the publically available opinion polls are high-quality data points. To iron-out the noise of an occasional bad poll, political analysts aggregate the polls to create a "poll-of-polls." The problem is that this method is mathematically the same as combining bad mortgages into securities. The idea is that each individual object (mortgage or poll) may be flawed, but if you get enough of them together, the problems will all average out and you have a very low risk of something bad happening.2 If there is a bias that is common to a large part of the data, then you are in real trouble. And why would there be a bias in election polls? For one, polling is not a science. It is an art. To extrapolate the results of an opinion survey of ~1,000 individuals to the general election of ~130 million people, polling professionals have to make turnout assumptions that are based partly on previous elections and partly on guesswork. This year, these assumptions are notoriously difficult to make as both candidates are extremely unpopular (Chart 2). This is bound to throw off pollsters' assumptions and may partially explain the regular gyrations that can be gleaned in Chart 1. For Secretary Hillary Clinton, the problem is compounded by the fact that she requires a high turnout to win. She needs the "Obama Coalition" of minorities and Millennial voters to show up as they did for President Barack Obama in 2008 and 2012. But we know that she struggled with the latter, with Senator Bernie Sanders picking up 70% of the youth vote in the Democratic primaries (Chart 3). If the 2016 turnout resembles the turnout from mid-term elections - which Republicans have generally won this century - then Trump could still have a chance. People may be lying: Another concern for Clinton is that she may be the 21st century Tom Bradley. Bradley was an African-American Mayor of Los Angeles who lost the 1982 California governor's race despite being ahead in the polls right up until election day. The "Bradley effect" theory goes that white voters lied when answering the polls in 1982 for fear of appearing racially prejudiced. Today, voters may be telling pollsters what they think is "politically correct," thus favoring Clinton in the polls. In the same vein - but ideologically opposite - the former Imperial Wizard of the Knights of the Ku Klux Klan, David Duke, outperformed expectations in both the 1996 U.S. Senate election and the 1999 special election for Louisiana's First Congressional District. He lost both elections, but he managed to garner double-digit support both times. More recently, the June 23 Brexit vote surprised markets. In our view, investors and betting markets underestimated Brexit largely in spite of polls, which had been close throughout the campaign stage (Charts 4 and 5). BCA's Geopolitical Strategy outlined the case for why the probability of Brexit was much higher than the market assumption as early as March.3 Our concerns began to manifest in the polls with the "Leave" camp comfortably ahead throughout June. And then, from June 16 (one week before the vote) to June 23, the "Stay" vote surged ahead in the polls, garnering a 4% lead the day before the election. This surge in the last week was clearly false, as the "Leave" camp won by a 3.8% margin, a 7.8% swing on the day of the election. So, what happened? The vertical line in Chart 5 shows the day that Member of Parliament Jo Cox was murdered by a British ultra-nationalist. Our guess is that the stunning political assassination - an extremely rare event in the U.K. - created a "Cox effect" in the Brexit polling. Those who were polled may have mourned for Cox, or resisted being associated with the extreme views of a self-professed neo-Nazi, yet they silently stuck to their legitimate concerns regarding EU membership on the day of the referendum. Chart 4Online Betting Got Brexit Wrong... Chart 5...So Did Prominent Opinion Polls The Brexit example illustrates that lying to pollsters is not something that only happens in the past. It has happened as recently as June. Given Donald Trump's controversial statements - and particularly his misogynist rants going back to 2005 - American voters may be lying to pollsters when it comes to their choice for president. Chart 6Media Narratives Are Cyclical Media narratives: As our geopolitical team has stressed throughout this election, the news media work through narratives (Chart 6). These narratives appear to have influenced polls, leading to regular gyrations in support levels for the two candidates. Will the media have another "comeback kid" narrative for Trump in store ahead of the election? It cannot be discounted. And if the polls tighten to the 0-3% range again, the turnout concerns and the "Bradley/Cox effect" from above could be enough to swing the election for Trump. Bottom Line: Clinton remains the favorite to win the election, but her probability of winning is closer to 65.5% than the 85% that appears to be "priced in the market." ...And Why He Will Not Win While we are not comfortable calling the election a "done deal," we do believe that Clinton is a favorite. The BCA Geopolitical Strategy quantitative model predicts that she has about a 65.5% probability of winning.4 And the team's qualitative analysis of Trump's electoral strategy suggests that the hurdles to his victory are considerable, particularly in swing states Virginia and Colorado. Before we introduce the quantitative and qualitative models that underpin our election forecast, let us address the above concerns about turnout and the "Bradley/Cox effect" head on. In our view, the polls are telling the truth. We concede that Trump's support level may be underestimated by approximately 3%, which would not be out of line with the last five presidential elections (Chart 7). However, a Clinton lead greater than ~3% the day of the election will be insurmountable for four reasons: GOP primary: It was not the polling that got Trump wrong during the Republican primary race, but the pundits. The polls were generally accurate, particularly those in the swing states where polls tend to be frequent and sophisticated (Chart 8). Polls only underestimated Trump by more than 3% in Illinois, Massachusetts, New York and Pennsylvania. Some of Trump's most controversial statements were made in late 2015 and early 2016 and yet they prompted no shame from his supporters when answering pollsters' questions. Turnout seesaw: Trump's strategy - which we dubbed "The Great White Hype" back in March - is a serious and mathematically viable electoral strategy.5 The effort focuses on boosting the GOP share and overall turnout of the white, blue-collar voter. The problem with this strategy, as executed by Trump, is that its effect could be a seesaw. Trump's rhetoric and policy proposals may appeal to less-educated, lower-income white voters, but may also reduce his support among well-educated, upper-income voters. This is a serious problem for Trump given that the 2012 exit polls indicate that Romney won college graduates by 4 points and voters earning $100k or above by 10 points. In other words, upper-income, well-educated voters are a key constituency of the Republican Party. And just as Clinton may have trouble getting Millennials and minorities to vote for her by the same margin as they did for Obama, Trump could be struggling to get key conservative constituencies out as well. Debates: All scientific polls taken after the debates have Hillary Clinton as a clear winner (Chart 9). This may seem surprising given the reaction of many pundits that Trump outperformed the very low expectations for him in the debates. Many analysts scored the debates close, but voters did not. Why? Because independent and undecided voters are just now tuning into the election and want to see candidates discuss serious policy issues and show leadership. Political science research shows that the direct influence of party identification decreases in presidential elections over time, but issues gain importance, especially after the presidential debates.6 As such, voters tuning into the debates were not discounting Trump's fiery rhetoric and behavior, they were appalled by it. We can't say we were surprised, as we have been showing Chart 10 to clients since February. Senate: If voters are hiding their true support level for Donald Trump, then their genuine preference should be revealed in Senate races where less controversial Republicans are contesting close elections. Instead, Republicans are on a path to lose four of their Senate seats, with another three in play (Democrats need four to take the Senate, assuming that Clinton wins the presidency, since Vice-President Tim Kaine would then cast the tie-breaking vote in that body). Democrats are ahead in Indiana, Illinois, Wisconsin, and Colorado. Nevada is also expected to stay blue. Missouri, New Hampshire, North Carolina, and Pennsylvania are all still in contention, despite the GOP incumbent advantage in all three. Bottom Line: Despite the challenges that this election presents - two highly disliked candidates, questions about turnout, and concerns about polling quality - we doubt that Donald Trump can surprise his poll numbers by more than ~3%. With Hillary Clinton up by 6.4% in the latest RealClearPolitics poll of polls, this means that Trump has to start rallying now if he is going to have a chance on November 8. What Do Our Quantitative & Qualitative Models Say? Our geopolitical team's quantitative model predicts that Hillary Clinton will win the election with 335 electoral votes. The model, built using historical macroeconomic and election data since 1980, has been projecting a strong Clinton victory for some time.7 It currently shows that Clinton already has 279 electoral votes from states where she has more than a 70% chance of winning (Chart 11). These results mean that even under the unlikely scenario in which the GOP wins all the remaining swing states (North Carolina, Arizona, Florida, Ohio, and Iowa), Clinton will still win the election, all other things being equal. Meanwhile, our qualitative model relies on testing Trump's electoral strategy - boosting the share of the white vote accruing to the GOP - in the real world. We concluded in March that Trump did have a path to victory, albeit a very narrow one. Our research showed that Trump's strategy is mathematically viable, at least in 2016 when the white share of the total population remains large enough. We specifically showed that Trump would only need to increase white voters' support by 1.7% and 2.9% in Florida and Ohio, respectively, to flip those states, which seems quite reasonable. We also pointed out that getting a 5.7% swing in Iowa could be feasible. On the other hand, we showed that "flipping" Midwest states like Michigan, Pennsylvania, and Wisconsin would require a very large swing of white voters in Trump's favor: 13.9%, 7.8%, and 8.1%, respectively. With those numbers, Trump would have to win nearly 70% of Michigan's white voters, 65% of Pennsylvania's, and 58% of Wisconsin's. Of the three, Wisconsin looks the most achievable. On the other hand, the GOP only managed to pick up 52% of the state's white share in 2004, the last time a Republican candidate for president won an actual majority of the popular vote since 1988. So, getting to 58% is a high bar given Wisconsin's recent electoral history. How did our qualitative model hold up in terms of state-by-state polling? It did really well! As we predicted, Trump has led the race or nearly led the race in Iowa, Florida, and Ohio (Chart 12). In Michigan, Pennsylvania, and Wisconsin, Clinton's lead has remained higher than 5% through most of the election cycle, even when the media narrative shifted against her (Chart 13). Chart 12The 'White Hype' Model Works Here Chart 13White Hype' Does Not Work Here If Trump were to win all the states that our White Hype model predicts as competitive, he would still be short of the necessary 270 electoral votes. Map 1 shows the ideal distribution of states for Trump, one that ignores the polls and assigns swing states to Trump or Clinton based on whether the White Hype model is feasible or not. Notice that the two remaining major states are Virginia and Colorado. For Trump to win this election, we believe that he needs to win one of the two (Colorado in combination with either Nevada or New Hampshire), in addition to all of Florida, Ohio, North Carolina, and Iowa. This is a tall order! Particularly given that his polling in Virginia and Colorado is poor (Chart 14). Chart 14Two Critical Swing States Bottom Line: BCA's Geopolitical Strategy quantitative and qualitative models both show that Hillary Clinton is a clear favorite to win the election, a view we have held since December 2015.8 Investment Implications: MNCs Vs. SMEs Our colleague Peter Berezin has already discussed the implications of a Trump victory: a stronger USD and a sell-off in stocks.9 We agree and would add that a rally in Treasurys would be likely in the event of a surprise Trump win (Chart 15). Chart 15Trump's Success Helps Safe-Haven Assets The rally in safe-haven assets would eventually give way, however, to a bear market in Treasurys as investors realized that Trump has no intention of controlling public spending or reining in the (already growing) budget deficit. Growth, and likely inflation, would surprise to the upside, allowing the Fed to hike rates beyond the 48 bps expected by the market through the end of 2018. We do not foresee that a Republican-held Congress would stand in Trump's way, despite the clear dislike between the Speaker of the House, Representative Paul Ryan, and Trump. Ryan would not go against a sitting president from the same party who just pulled off a revolutionary election. The entire House will face re-election in 2018 and moderate Republicans will be wary of standing up to Trump, lest he campaign against them in GOP primaries in a short two years. Investors are putting way too much faith in America's checks-and-balances to keep Trump from enacting his policies, at least in the short term. These are constitutional, legal, and technical checks, and political expediency often overrules all three. In case of a Clinton win, we would expect the House to remain controlled by the GOP. There are only about 38 truly competitive electoral districts in this race, according to The Cook Political Report.10 Given that the Republicans have a 60-seat majority in the House, a Democratic takeover would require Democratic candidates to defeat Republican Representatives in 30 out of 38 competitive districts. At best, this means that the current, market- bullish status quo of divided government will continue. With the House remaining in Republican hands, and Democrats clinging to a potential razor-thin control of the Senate (vulnerable to a post-Trump Republican comeback in 2018), the Clinton White House would be constrained on some of its most left-leaning policies.11 And what are the chances of cooperation on modest reforms? We think they are actually quite good. Unlike Obama, Clinton's victory will not be a popular sweep. She will not control Congress, she will likely receive less than 50% of the popular vote (due to the presence of two notable third-party candidates), and she will be the first candidate ever elected that has more voters saying they dislike her than like her. Therefore, the odds are slim that Clinton will come to power with the same level of confidence and agenda-setting vision as Obama did in 2008. Instead, we see two potential avenues for modest cooperation with the GOP-controlled House: Chart 16Corporate Taxes Have Bottomed Corporate tax reform: It is unlikely that we will see reform that lowers the already historically-low effective tax rates (Chart 16). However, broadening the tax base by closing various loopholes could be feasible. This will hurt S&P 500 multi-national corporations that have been able to lobby for special treatment over the past three decades. However, it will benefit America's SMEs, which are the backbone of employment and growth. Fiscal spending: Paul Ryan and moderate Republicans understand that there is a paradigm shift in America and that the median voter is moving to the left.12 After all, Donald Trump won the GOP primary with an unorthodox economic message that combined both left- and right-wing economic policies. As such, we would expect House Republicans to give in to a modest infrastructure spending plan from Clinton, in exchange for corporate tax reform. Even a modest plan could make a substantive difference for the economy given the high fiscal multipliers of infrastructure spending in an economy with low interest rates. This in turn would allow the Fed to surprise the markets with more than two rate hikes by the end of 2018 and thus sustain the USD bull market. If there is one trend that we are certain will end with the 2016 U.S. election, it is the dominance of American economic policy by the S&P 500, or perhaps the S&P 100. What Trump and Senator Bernie Sanders have shown is that challenging for the presidency no longer requires a cozy relationship with either Wall Street or the large multinational corporations (MNCs). We therefore do not expect a Clinton-Ryan coalition to care as much about the concerns of America's large corporations as otherwise might be the case. Policies that lead to higher effective corporate tax rates on major S&P 500 corporations, a dollar bull market, and higher wages are likely over the course of the next four years. The political pendulum is shifting in the U.S. and it should marginally favor growth, inflation, the USD, and SMEs.13 Marko Papic, Senior Vice President Geopolitical Strategy marko@bcaresearch.com 1 Please see FiveThirtyEight, "Who Will Win The Presidency?" dated October 20, 2016, available at fiverthirtyeight.com. 2 "You mean like the 2008 Global Financial Crisis?" Yes. Like that. 3 Please see BCA Geopolitical Strategy Special Report, "With Or Without You: The U.K. And The EU," dated March 17, 2016, available at gps.bcaresearch.com. 4 Please see BCA Geopolitical Strategy Special Report, "U.S. Election: Final Forecast & Implications," dated October 12, 2016, available at gps.bcaresearch.com. 5 Please see BCA Geopolitical Strategy and Global Investment Strategy Special Report, "U.S. Election: The Great White Hype," dated March 9, 2016, available at gps.bcaresearch.com. 6 Please see Andreas Graefe, "Issues and Leader Voting in U.S. Presidential Elections,"Electoral Studies 32:4 (2013), pp.644-657. 7 For the assumptions underpinning our model, we encourage clients to read BCA Geopolitical Strategy Special Report, "U.S. Election: Final Forecast & Implications," dated October 12, 2016, available at gps.bcaresearch.com. 8 Please see The Bank Credit Analyst Strategy Outlook, "Stuck In A Rut," dated December 17, 2015, available at bca.bcaresearch.com. 9 Please see BCA Global Investment Strategy Special Report, "Three (New) Controversial Calls," dated September 30, 2016, available at gis.bcaresearch.com. 10 Please see "House: Recent Updates," accessed October 20, 2016, available at cookpolitical.com 11 We believe that it will be very difficult, if not impossible, for the Democrats to retain a razor-thin majority in the Senate if they get one in November. First, Democrats will have to defend 25 Senate seats (including two allied independent seats) out of 33 in contention in 2018. Second, Democrats always see a drop-off in voter turnout and enthusiasm in mid-term elections. 12 Please see BCA Geopolitical Strategy Monthly Report, "Introducing: The Median Voter Theory," dated June 8, 2016, available at gps.bcaresearch.com. 13 Please see BCA Geopolitical Strategy Monthly Report, "King Dollar: The Agent Of Righteous Retribution," dated October 12, 2016, available at gps.bcaresearch.com. Strategy & Market Trends Tactical Trades Strategic Recommendations Closed Trades

Special Report

Highlights Muni Credit Cycle: The reading from our Municipal Health Monitor supports low Muni/Treasury yield ratios for now, but will become less supportive near the end of 2017. This is consistent with historical lags between the muni and corporate credit cycles. Issuance: State & local government investment spending will increase in 2017, as will muni issuance for new capital. Pensions: The pension funding problem will only get worse in the coming years. Credit ratings do not adequately reflect the risk from under-funded pensions. Election: Muni/Treasury yield ratios have not yet discounted Donald Trump's recent plunge in the polls. Maintain an overweight allocation to municipal bonds for the time being, but stand prepared to gradually reduce exposure as the muni credit cycle starts to turn in late 2017. Feature The financial crisis marked a major inflection point in the municipal bond market. Not only did the economic fall-out from the housing crash lead to historically large state & local government budget gaps, but the end of bond insurance and a growing realization that municipal default is possible have focused investor attention on credit quality more than ever before. In this Special Report we zero-in on the Municipal/Treasury (M/T) yield ratio.1 We identify its major short-run and long-run drivers, and assess where it is headed in the context of the municipal bond credit cycle. The Longer-Run Outlook For Yield Ratios An important input to our understanding of the municipal credit cycle is our Muni Health Monitor (MHM). The MHM is a composite of eight indicators of state & local government budget health. Full details of the indicator and its components can be found in the Appendix to this report. The MHM has an excellent track record of signaling the major inflection points in muni ratings migration (Chart 1). We observe that the MHM bottomed in 2006, one year before the previous trough in ratings migration. The MHM also crossed into "deteriorating health" territory six months before municipal downgrades started to outpace upgrades in 2008. More recently, the MHM crossed back into "improving health" territory in Q4 2012. Muni ratings migration also peaked in Q4 2012 and upgrades began outpacing downgrades in Q4 2014. Chart 1The Municipal Health Monitor Leads Ratings We pay attention to the trends in muni ratings because ratings and state & local government net borrowing explain more than 50% of the variation in the average M/T yield ratio since 1997 (Chart 2). Further, increased investor focus on the creditworthiness of municipal issues has made the yield ratio even more responsive to ratings and net borrowing since the Great Recession. So where are we currently situated in the muni credit cycle? The MHM remains in "improving health" territory, but appears to have entered an extended bottoming-out phase. Given the re-leveraging that has already occurred in the corporate sector, it would be extremely unusual for the MHM to improve further during this cycle. In fact, our Corporate Health Monitor tends to lead the MHM by about two years (Chart 3). This squares with what we know about the behavior of state & local governments throughout the economic cycle. Chart 2The Muni Credit Cycle Illustrated I Chart 3The Muni Credit Cycle Illustrated II Typically, the corporate sector will increase debt loads when times are good and will then be forced to de-lever when the economy enters recession, profits contract and those debt loads become unsustainable. State & local government budget gaps, on the other hand, will tend to narrow during an economic recovery as rapid income growth translates into increased tax revenue. It is only during a recession that state & local government budget gaps widen, since tax revenue plummets while expenditure growth - particularly for social benefits - remains firm. The end result is that the municipal credit cycle tends to lag the corporate credit cycle. This is also apparent in the ratings data (Chart 3, bottom panel), which suggest that we should expect to see muni downgrades (and hence yield ratios) head higher near the end of next year. The typical lag between the corporate credit cycle and the municipal credit cycle suggests that M/T yield ratios should remain well behaved until late-2017, and then begin to move higher. However, the extraordinary length of the current economic recovery gives us some cause to believe that the lags in this cycle may be somewhat longer. We turn to a macro analysis of net state & local government borrowing to shed some further light on this issue. Net borrowing is simply the difference between revenues and expenditures. On the revenue side of the ledger, state & local governments have already seen a significant deceleration in tax receipts during the past year (Chart 4). Every source of tax revenue - except for property taxes - has slowed alongside what has been disappointing overall economic growth so far in 2016. While a return to the 10% revenue growth that was seen in the mid-2000s is unlikely, we expect most of the recent deceleration will soon be reversed. Aggregate weekly hours bounced sharply in September (Chart 5), and federal income tax withholdings also continue to grow rapidly. Both indicators suggest that income growth will be stronger during the next few months, which will support state & local tax receipts. On the expenditures side, while spending on social benefit programs has increased, state & local governments have largely dealt with budget gaps by cutting back severely on discretionary spending (Chart 6). Investment spending has also collapsed and, as a result, gross municipal bond issuance has been dominated by refinancing (Chart 6, bottom two panels). Chart 4S&L Government Revenue Chart 5Income Growth Will Rebound Chart 6S&L Government Expenditures This could all be about to change. Both U.S. Presidential candidates have prioritized infrastructure spending as part of their platforms. Hillary Clinton plans to increase infrastructure spending by $500 billion. This consists of $250 billion of federal infrastructure spending over the next five years and $25 billion of seed money to create a national infrastructure bank. The bank would also accept an additional $225 billion in direct loans. Clinton's plan would also bring back the Build America Bonds (BABs) program. Donald Trump has also expressed a desire to invest heavily in infrastructure, and has floated figures in the range of $1 trillion, although he has been less specific about the details. Historically, about 70% of public investment has occurred at the state & local government level (Chart 7). This suggests that if infrastructure spending became a priority it would lead to a large increase in state & local government investment and hence municipal bond issuance. However, with Clinton's plan it is still unclear whether the bulk of infrastructure spending would be financed through the Treasury market or the muni market. Certainly, to the extent that increased spending is financed through the BABs program, then tax-exempt muni issuance would not be impacted. In our view, state & local government investment spending will head higher in 2017 even without any support from the new President. The need for state & local governments to invest in infrastructure has been evident for some time, but only recently have budgets become healthy enough for governments to consider it. There is a strong correlation between state & local government investment spending and the net percentage of states with a total balance (general fund plus rainy day fund) that exceeds 5% of expenditures (Chart 8). This figure has just recently moved into positive territory and, not coincidentally, more than $200 billion worth of infrastructure spending will be on ballots requesting voter approval in November.2 Chart 7State & Local Government ##br##Drives Investment Chart 8Healthy Enough##br## To Invest The combination of resilient, but not surging, revenue growth and increased investment spending in 2017 is consistent with the idea that the muni credit cycle will follow the lead of the corporate cycle and start to turn near the end of next year. Bottom Line: The reading from our Municipal Health Monitor supports low Muni/Treasury yield ratios for now, but will become less supportive near the end of 2017. This is consistent with historical lags between the muni and corporate credit cycles. The Pension Problem Of course, the elephant in the room with regards to the long-run outlook for municipal credit quality is pensions. So far pensions have only entered our discussion of the muni credit cycle tangentially, since the pension funded ratio is a component of the MHM (see Appendix). However, large unfunded pension liabilities - should they persist - have the potential to be severely destabilizing for the muni market at some point in the future. According to the U.S. National Accounts, aggregate defined benefit pension entitlements at the state & local government level total $5.6 trillion, only 65% of which are currently funded by assets. However, this aggregate figure masks large divergences between a few municipalities with unsustainable pension liabilities and the majority of municipalities where pension liabilities are probably manageable. Chart 9Low Returns Put Pressure On Pensions In a recent report,3 the Center for Retirement Research at Boston College found that 36 states should be able to fund their existing liabilities by making annual payments that total less than 15% of revenue. However, five states - Illinois, New Jersey, Connecticut, Hawaii and Kentucky - require annual payments in excess of 25% of revenue. The breakdown is found to be similar at the city level, where pension costs were found to be manageable for the majority of cities, although Chicago, Detroit, San Jose, Miami, Houston, Baltimore, Wichita and Portland all face annual pension costs that exceed 40% of revenue. Unfortunately, while the pension situations of most municipalities are currently manageable, they are only likely to get worse. Changes in the aggregate pension funded ratio closely track returns from a portfolio that is 50% invested in the S&P 500 and 50% invested in the Barclays Treasury index (Chart 9). Based on current equity valuations, it is probably only reasonable to expect 6% annual nominal returns from the equity market during the next 10 years,4 and the 10-year Treasury yield suggests that 1.8% is a reasonable expectation for annual nominal Treasury returns. Taken together, annual nominal investment returns from a 50/50 portfolio during the next decade could be close to 4%, far below the historical average of 8.9% and also below the 7.6% average return assumed by state & local pension plans in 2014. The two main points are that: The pension problem is likely to get worse, not better Given that large under-funded pensions are concentrated in only a few states, inter-state muni allocations are very important On this second point, we observe that states with lower pension funded ratios have higher General Obligation (GO) bond yields (Chart 10), and also that not all of the difference is reflected in credit ratings. We ran a cross-sectional regression of GO bond yields against credit rating and found that a correlation remains between the residual from that regression and the pension funded ratio (Chart 11). In other words, credit rating does not adequately control for the risk presented by under-funded pensions. Chart 10Municipal Bond Yields Vs. Pension Funded Ratios Chart 11Municipal Bond Yields Vs. Pension Funded Ratios: Controlling For Credit Rating Bottom Line: The pension funding problem will only get worse in the coming years. Credit ratings do not adequately reflect the risk from under-funded pensions. The Short-Run Outlook For Yield Ratios So far we have discussed the muni credit cycle and noted that M/T yield ratios should begin to move higher on a sustained basis at some point near the end of 2017. However, the near-term drivers of M/T yield ratios suggest that an overweight allocation to municipal bonds remains appropriate for the time being. We have found that the bulk of near-term volatility in M/T ratios can be explained by four factors (Chart 12): The Global Policy Uncertainty Index5 Gross municipal bond issuance Net municipal mutual fund flows Ratings migration The Brexit shock to policy uncertainty has now mostly been reversed. Meanwhile, our Muni Excess Supply Indicator (Chart 12, panel 4) shows that gross issuance has been outpacing fund inflows of late. This should put upward pressure on yield ratios, although this pressure has been largely offset by still supportive ratings migration (Chart 12, bottom panel). Considering all factors, this short-term model shows that the average M/T yield ratio is close to fair value. A reading close to fair value is consistent with muni returns that should exceed those from duration-equivalent Treasuries most of the time (Table 1), even before adjusting for the muni tax advantage. In fact, Table 1 shows that the odds of muni underperformance only really increase once the M/T ratio appears more than one half standard deviation expensive on our model. Chart 12A Short-Term Muni Model Table 1Municipal Bond Excess Returns* Based On Fair Value Model** Residual: 2010 - 2016 The other near-term factor that supports a continued overweight allocation to municipal debt is the prospect of a Clinton victory in next month's election. Since the beginning of the year, the average M/T ratio has closely tracked the probability of a Trump election victory (Chart 13). The reasoning is entirely logical. Trump has promised large tax cuts for the highest earners. Such tax cuts would significantly de-value the tax advantage of municipal bonds and pressure yield ratios higher. In contrast, Clinton promises to raise taxes on high income individuals. This would make the tax advantage of municipal debt more valuable, and pressure yield ratios lower. Chart 13Trump Is Bad For Yield Ratios The average M/T yield ratio has not yet discounted Trump's recent plunge in the polls. This argues for the maintenance of an overweight allocation to municipal debt in the near term. Bottom Line: M/T yield ratios appear fairly valued in the near-term, and have not yet discounted Donald Trump's recent plunge in the polls. Maintain an overweight allocation to municipal bonds for the time being, but stand prepared to gradually reduce exposure as the muni credit cycle starts to turn in late 2017. Ryan Swift, Vice President U.S. Bond Strategy rswift@bcaresearch.com Alex Wang, Research Analyst alexw@bcaresearch.com Appendix: The BCA Municipal Health Monitor The BCA Municipal Health Monitor is an equal-weighted composite of eight indicators meant to quantify trends in state & local government budget gaps and debt service capability. The components consist entirely of data that are publicly available from the Bureau of Economic Analysis, Federal Reserve, Bureau of Labor Statistics and the National Association of State Budget Officers. The eight components are described below, and shown graphically in Charts A1 & A2. Chart A1Muni Health Monitor Components I Chart A2Muni Health Monitor Components II Leverage: The ratio of total state & local government liabilities (excluding unfunded pension liabilities) to total financial assets. Interest Coverage: State & local government current budget surplus (excluding interest expenditures) divided by interest expenditures. The current surplus is calculated as the difference between current revenues and current expenditures (i.e. investment spending is excluded). Pension Funded Ratio: Total assets of state & local government pension funds divided by total pension liabilities. Revenue: State & local government current revenue in nominal terms, as a deviation from its 18-quarter trend. Surplus Margin: State & local government current budget surplus as a % of current revenue. Liquidity: State & local government total financial assets less short-term liabilities, as a % of total financial assets. Employment Growth: Year-over-year % change in state and local government employment. Total Balance: Aggregate state government total year-end balance. The total balance is the general fund balance plus the rainy day fund, as a % of total expenditures. 1 The average M/T yield ratio shown in this report is calculated by taking an equal-weighted average of M/T yield ratios for 2-year, 5-year, 10-year and 30-year maturities. For each maturity point the yield ratio is calculated as the ratio between the Bloomberg Fair Value Aaa Municipal bond yield and the Federal Reserve's constant maturity Treasury yield. 2 http://www.bloomberg.com/news/articles/2016-09-13/mega-deals-lead-ballo… 3 http://crr.bc.edu/briefs/will-pensions-and-opebs-break-state-and-local-… 4 Please see Global Investment Strategy Special Report, "Global Equity Valuations: Risks And Opportunities", dated July 1, 2016, available at gis.bcaresearch.com 5 The index was created by Professors Scott Baker, Nick Bloom and Steven Davis and is driven by the number of times terms related to economic and policy uncertainty are found in newspaper articles. Full details of the methodology are available at www.policyuncertainty.com

Highlights Our protector portfolio is a combination of assets that have a low or negative correlation with equities that give investors some downside protection. Replacing cash and/or Treasuries with our protector portfolio in 60-30-10 or 60-40 benchmark portfolios would have produced superior returns since 2011. We continue to advocate allocating investments to our protector portfolio in the near term as it represents an effective hedge against immediate risks such as a negative market reaction to the upcoming elections and/or disappointing third quarter profits. Feature Both equities and bonds are under pressure, as a higher likelihood of a December interest rate hike is beginning to be priced in at the same time as nervousness about Q3 earnings results has intensified. This confluence of factors - less liquidity and earnings disappointment - has been the central argument of our defensive portfolio stance for some time: any handoff from liquidity to growth would be shaky, and potentially premature. Indeed, as we wrote in the September 26 Weekly Report, liquidity conditions will largely remain favorable for risk assets for some time because even with a December rate hike, interest rates are well below equilibrium, i.e. are not restrictive. However, equity investors will suffer through bouts of earnings disappointments, similar to the chronic disappointment in GDP growth. As we show in Chart 1, throughout the economic recovery, expectations for economic growth have been revised lower and are only now finally in line with what we expect is close to reality. As highlighted in last week's report, investors' expectations about earnings are most likely to undergo the same fate because profit margins will remain a lasting headwind: investors have not yet adjusted to this new reality (Chart 2). That will hold equity gains to low single digits, at best. Chart 1Years Of One-Way (Down) Revisions Chart 2Earnings Set To Disappoint? Overall, our view is that the economic backdrop is stable as there are low odds of a recession-inducing monetary tightening occurring, and we do not see any other negative shocks that are concerning enough to trigger a recession. Still, above and beyond our worry about profit disappointments, many client queries are currently focused on U.S. election risks. On September 26, we warned of market volatility leading up to the election, since investors may continue to assign too low odds of a Trump Presidential win. However, we would expect markets to quickly recover - at least until Trump reveals his true policy colors. We took a page from the market reaction to Brexit as a possible guideline to the outcome of Trump winning the election, i.e. the election is ultimately won by a non-status quo candidate. Investors will recall that the post-vote U.K. equity market reaction to Brexit was short-lived but savage. However, the uncertainty around the upheaval of institutions and structures in the euro area and the U.K. are far greater than the election of a non-conformist U.S. President within an institutionally sound system with checks and balances. All of that said, we recognize that we could be wrong and that the U.S. election has taken over the pole position on investors' list of concerns. More specifically, investors are worried about negative financial market fallout from a Trump win.1 So, how should investors hedge the downside risk of these election results? And for that matter, what about other near-term risks? Protector Portfolio Explained This publication has been advocating for some time that investors hold some portion of their capital in a protector portfolio (currently a combination of TIPS, gold and the U.S. dollar). The goal is to find assets with a low or negative correlation to U.S. equities and offer a measure of protection against a steep selloff in stocks. As Chart 3 shows, a portfolio of 60/30/10, where 10% is placed in the protector portfolio, would have outperformed a traditional 60/30/10 allocation in which the 10% is held in straight cash since 2011 (in a ZIRP world). A 60/40 allocation where 40% is placed in the protector portfolio also beats a 60/40 stock/Treasury allocation since 2011. Chart 3Protector Portfolio Enhances Performance ##br## Since 2011 Chart 4Protector Components Are ##br## Negatively Correlated With S&P 500 The three assets included in our protector portfolio were chosen with specific risks in mind: USD: As the main global reserve currency, the U.S. dollar benefits when global risk aversion is on the rise. Admittedly, when fears have emanated from U.S. soil, the dollar has performed less well compared to other safe-haven assets, such as the Swiss franc and/or Swiss bonds. Nonetheless, for U.S. investors, investing in one's home currency can provide a natural hedge/advantage. In Chart 4, we show the one-year correlation between USD and S&P 500 equity returns. Since 2009, the correlation has been negative and the implication is that by holding USD, investors are already implicitly defensive. Gold: Gold traditionally does well in times of extreme geopolitical uncertainty and also as a hedge against inflation. More recently, gold has done less well as a hedge because the negative correlation between equity prices and gold broke down from 2011 until earlier this year (Chart 4). Gold has once again become negatively correlated with equity prices and we believe it will be an effective safe-haven asset should inflation become a concern. TIPS: Both 10-year TIPS and nominal Treasuries are negatively correlated with U.S. equity returns and both provide some measure of insurance in risk-off periods/phases of economic disappointment. Nonetheless, we prefer TIPS at the moment since they offer a measure of protection against a back-up in inflation expectations (also Chart 4). In sum, our protector portfolio is a combination of assets that are uncorrelated enough with equities to give investors some protection against a range of downside risks. Protector Portfolio: But Beware Buy And Hold Chart 5Protector Buy And Hold Will Not Work As Chart 2 has shown, our protector portfolio has outperformed both a 60-30-10 and 60-40 portfolio in recent years. However, longer -term performance has been less outstanding (Chart 5). Indeed, adding a constant proportion of safe-haven assets to a balanced portfolio over an extended period underperforms the balanced portfolio benchmark for long stretches of time: there are non-negligible costs associated with holding safe-haven assets over prolonged periods. The bottom line is that timing plays a critical part in investing in safe-haven assets. Owning a fixed share of protector portfolio assets over long horizons will not beat a traditional buy and hold strategy, although superior returns over cash offer a compelling case in a NIRP world. We continue to recommend that investors hedge against downside risk in the form of the protector portfolio - or simply by choosing the safe haven that most closely corresponds as a hedge to the specific risk at hand. However, it is important to know that safe-haven assets fall in and out of favor through time and the protector portfolio will at some point no longer be justified, and/or its components will need to be adjusted. For example, only after 2000 did Treasuries start providing a good hedge against equity corrections. The contrary is true for gold - it acted as one of the most secure investments during corrections until that time, but then became correlated with S&P 500 total returns from 2012-early 2016. That said, gold's coefficient has turned negative again, and it should be viewed as an all-weather safe haven, especially if deflation risks begin to dissipate. The Most Relevant Safe Haven In Case Of A Policy Mistake Chart 6Fed Policy Mistake? Buy Protector Portfolio As we wrote above, our base investment case is that the prospect of less liquidity and the risk of an earnings disappointment mean that investors should keep a defensive portfolio stance and be prepared for pullbacks in equities in the single digits. However, the Minutes of the latest FOMC meeting highlight that a fairly low threshold has been set for a December interest rate rise. If financial market participants interpret incoming economic information more bearishly than the Fed, then a December rate hike risks being perceived by investors as a policy mistake. Under this scenario, risk assets could be set for a much greater fall, buoying the case for further portfolio insurance. Which safe havens will outperform? We take our cue from the market reaction to the December, 2015 rate hike. In that episode, equity prices fell 12%. The protector portfolio in its current configuration2 increased 10%. The bulk of the appreciation was due to a strong run in gold prices (surely helped in part by massive woes in China) and TIPS (Chart 6). We believe that this basket of assets would once again offer an important buffer against equity losses associated with a policy mistake. The Most Relevant Safe Haven For A Trump Win If a Trump win triggers a correction in risk assets, we would expect the U.S. dollar to rally due to Trump policy uncertainty and heightened geopolitical risk. We noted above that USD does not always rally when a stress event occurs on U.S. soil. However, in the past several weeks, the performance of the dollar as well as Treasury yields has been linked to Trump's probability of winning the election. Whenever the odds of a Trump presidency rise, these risk-off assets have appreciated. And The Most Relevant Lessons From The Election Cycle This month's Geopolitical Strategy Special Report 3 provides a final forecast and implications for the elections. As we note above, we agree that a Trump win is a red herring in terms of the key issues investors face. But we also agree with our geopolitical strategists that there are several important lessons from the election cycle that may have long term ramifications for investors. Below, we highlight the most relevant for financial market participants: The median voter has moved to the left on economic policy. Trump's victory over an army of seasoned, relatively orthodox GOP contenders in the primary exposed the fact that the party's grassroots voters no longer care deeply about fiscal austerity and no longer wish to tolerate the corporate incentive for importing cheap labor. Similarly, demographic trends favoring millennials and minorities (who tend to vote left on economic policies), portends a shift by which the GOP attempts to capture left-leaning voters. Fiscal conservatism (and social conservatism, for that matter) will have less to show by way of official party machinery. The 2016 election campaign has amplified the notion that the news media works in narratives. These narratives work as a filter that preempts and distorts the presentation and, to some extent, reception of facts. This phenomenon was influential in Trump's rise - the first "Twitter" candidacy - as well as his recent decline. Investors cannot be too wary of what the mainstream press or financial "smart money" says about any particular political trend or event. It is essential to separate the wheat from the chaff by using empirics and looking at macro and structural factors to identify the constraints rather than the preferences of candidates or politicians. U.S. Economy: Neither Hot Nor Cold The NFIB survey of small business survey ranks as one of our preferred indicators of U.S. business confidence. The employment related indicators serve as a key input into our payroll model; questions about the pricing environment often provide a good leading/coincident gauge about inflation trends, and; as Chart 7 shows, the labor cost versus pricing series provides an excellent leading indicator for the profit margin outlook. The latter remains in a downtrend, reinforcing our message that profit margins will remain a headwind to earnings growth for still some time. Overall, small business optimism has been generally flat this year, after peaking in late 2014. It is somewhat discouraging that "demand" as a most important problem is no longer falling. Consumption has been one of the more robust areas of growth in the past several years and we expect consumption to continue to outshine other areas of the economy. However, even here, the data should be monitored closely. Chart 7Small Business Concerns (Part 1) Chart 8Small Business Concerns (Part 2) Retail sales (excluding gasoline and autos) growth has been slowing throughout 2016 and September data did not buck this trend (Chart 8). Results among retailers varied substantially, with growth strongest at building supply stores, sporting goods stores, vehicle dealers and furniture stores. Laggards include electronics and appliance stores - segments that are still under siege from falling prices. The bottom line is that in aggregate, consumption is holding up reasonably well and should continue to do so, as long as employment gains and modest wage growth remain intact. Stay tuned. Lenka Martinek Vice President, U.S. Investment Strategy lenka@bcaresearch.com 1 Our Geopolitical Strategy service concurs that a Trump win is a red herring, i.e. is unlikely to occur and is a distraction from more relevant issues. For more insight, please see Geopolitical Strategy Monthly Report "King Dollar: The Agent Of Righteous Retribution", dated October, 2016, available at gps.bcaresearch.com 2 At the time, the protector portfolio performed slightly less well, as 30-year government bonds were used instead of TIPS. 3 Please see Geopolitical Strategy Special Report "U.S. Election: Final Forecast & Implications", dated October 12, 2016, available at gps.bcaresearch.com Market Calls

Hillary Clinton has a 65% chance of winning the election; she receives 334 electoral college votes according to our model. Trump still requires an exogenous shock to win. Meanwhile, the USD is poised to rally - and leftward-moving policymakers will applaud its redistributive effects while MNCs suffer the consequences.

Special Report

Hillary Clinton has a 65% chance of winning the election; she receives 334 electoral college votes according to our model. Trump still requires an exogenous shock to win. Meanwhile, the USD is poised to rally - and leftward-moving policymakers will applaud its redistributive effects while MNCs suffer the consequences.

Our Treasury yield fair value model suggests that the 10-year Treasury yield has an additional +19bps of upside. Stay at below benchmark duration.

As the U.S. median voter is shifting to the left, redistributive policy could come into play. A strong dollar helps to achieve this goal as it results in a bigger share of labor income in the economy. EM and commodity currencies could bear the brunt of the pain. Favor the euro on its crosses. Stay short CAD/NOK, but tighten stops.

Special Report

This week's <i>Special Report</i> looks at the three controversial predictions that I made at this year's <i>BCA New York Investment Conference</i>.

This month's <i>Special Report<i/> looks at the Fed's policy options in the event that there is a negative economic shock while the policy rate is still very depressed. The Fed's "Plan A" is more QE and forward guidance, which are not up to the task. There is no "Plan B", which means that risk assets will be hit hard during the next downturn.

The monetary policy sweet spot won't end for equities until interest rates climb above the equilibrium rate - that won't happen even with a rate hike in December. But even though equity markets will continue to cheer a go-slow Fed in the short run, the sustainability of these gains will be dubious until a healthier environment for earnings takes hold.