Rumor is that taxes might fall and interest rates might rise further, leading one to ask, “Now, what do I do?” We advise: that depends on you, how your current portfolio is constructed and your definition of success. In this Perspectives, we assess the potential impact of a decline in corporate tax rates and rising interest rates on asset allocation and earned investment income, relying on industry aggregates as a representative company proxy.
We address asset allocation in an enterprise context, contrasting “optimal” asset allocations in two scenarios: using existing tax rates, and applying an alternative tax structure suggested by current political rhetoric. The asset allocation addresses the role of tax-preferenced securities such as tax-exempt municipal bonds and DRD eligible equities. In each scenario, the optimal asset allocation is the portfolio construction that provides the highest risk-adjusted after-tax return on equity (subject to estimates of initial risk tolerance).
Next, we assess the potential impact of alternative tax rates on earned investment income as interest rates are increased. In this assessment, insurance underwriting results do not vary, but the outcomes are tax effected at the changing tax rates. The “base” case for asset allocation (current tax rates) and earned investment income projections (primarily due to prospective rates and spreads) were presented in a previous issue of Perspectives.1
The key findings of our review are:
• The impact of alternative tax rates on asset allocation is influenced significantly by the return metric: book income versus total return (what matters most to you?).
• The impact of alternative tax rates is also dependent on the embedded yield-term structure and the assumed prospective term structure of taxable and tax-exempt capital market opportunities.
• Asset allocation will be less dependent on underwriting outcomes in a low-tax-rate environment.
• Rising interest rates will benefit all insurers’ earned investment income, although reported interim total return will lag book income due to price depreciation (do you really care?).2
The asset allocation review takes several sequential steps: first, establish a base case “optimal” allocation predicated on a set of stated assumptions; second, estimate the impact of a change in tax rates on the base case; and third, repeat the estimation assuming a change in interest rates (levels and spreads). For the second and third steps we assume tax-preferenced proration is eliminated.
Table I below displays an initial optimal property and casualty (P/C) industry asset allocation based upon:
• Year-end 2015 cusip detail holdings and premium to surplus and investment to surplus leverage3
• No change in interest rate levels or spreads
• A combined ratio of 97 (3% product margin) and historic underwriting and investment volatility
• An enterprise capital loss threshold of a 99.5 tail-value-at-risk (T-VaR)
• Current tax regime rates and proration rules
As shown in Table I, the industry’s estimated after-tax return on equity is 7.82% and the 99.5 enterprise T-VaR is 31.56% of capital (the expected loss of capital if the initial VaR threshold is exceeded).4 A more optimal asset allocation, while not exceeding the current estimated T-VaR, results in an after-tax return on equity of 8.42%, an increase of 60 bps after-tax.
The asset allocation changes are also shown in Table I. These include: a reduction in U.S. Governments and agencies and investment grade credit; an increase in municipal bonds, structured securities, preferreds, equities and alternatives; a .57 year duration increase (driven by municipal bonds and preferreds); and a constant average credit rating of AA-. Note that preferreds have an assumed rating of BB.
Embedded within the equities and alternatives allocation is an increase in high dividend equities to 2.5%. The resulting tax-preference holdings in the “optimal” portfolio totals 37.5% (municipals, preferreds and high dividend equity). The impact upon the “optimal” asset allocation of reducing tax rates to 20% (and eliminating tax-preferenced proration) is shown in Table II.
The initial after-tax total return on equity in the 20% tax rate environment is 9.16% or 134 bps higher than when taxes are at 35%. Approximately 100 bps of the differential is due to the impact of the tax rate reduction on invested assets (45 bps of invested assets’ after-tax total return differential times 2.2 of investment leverage). The remaining 34 bps is due to the impact of the tax rate change on underwriting results. Also note, the estimated T-VaR in the 20% tax environment is lower due to the higher after-tax income. (This begs a question for another day’s discussion: How should taxes matter?). For comparative purposes, we assume the estimated current T-VaR of 31.56% of capital remains the targeted threshold.
The “optimal” asset allocation is quite different. Vis-à-vis the current environment, tax- preferenced holdings are eschewed when rates fall to 20%: municipals, preferreds and dividend stocks have a much less attractive after-tax return and risk profile. U.S. Governments and agencies remain sub-optimal choices. Investment grade credit regains a portion of its lost allocation in the 35% tax environment, while structured securities further brighten their already glowing halo. Duration increases .5 years, as long dated tax-exempts provide less relative value at significantly lower tax rates. Average credit remains constant.
The asset allocations shown in Tables I and II are very sensitive to the current portfolio’s fixed income term structure, the fixed income reinvestment opportunities’ term structure and equity returns. Preferential tax treatment is an additional consideration.
Chart 1 shows year-end 2015 taxable and tax-exempt after-tax book yields of industry fixed income holdings and current market yields for single A-rated corporate bonds and double AA-rated tax-exempt bonds5. The data shown assume a 35% tax rate and proration.
The double AA tax-exempt tax book yields after-tax are greater than single A taxable market (book) yields for (nearly) all maturity periods. Regardless of prospective tax rates, these current exempt holdings would not be replaced but would be allowed to run-off. However, if interest rates were to increase and the return metrics were total return rather than book yield, then the certitude of this statement might diminish as the severity of rising rates increases. Table III displays the potential impact of tax rate changes and a modest rise in rates.
Columns one and two were shown in Table II. Total return and income return on assets are after-tax metrics. The sole cause of the difference between their amounts is the assumed price appreciation of equity holdings (bonds are being carried at amortized cost). In columns three and four, the prospective price depreciation of fixed income assets is reflected in total return on assets, as yields are assumed to rise. The initial income return is unchanged; however, the initial total return of 2.01% is 84 bps less than when bonds are valued at book value.
At a 35% tax rate, the allocation increases for tax-preferenced assets (municipals, preferreds and high dividend equity) whether or not fixed income price depreciation is taken into account. In Table III columns five and six the tax rate is 20%, proration is eliminated (municipals and eligible dividends are fully exempt) and fixed income price depreciation is disregarded; and the tax preferenced assets’ allocation, in total, is unchanged.
In columns 7 and 8 of Table III, fixed income price depreciation (rising rate environment) is taken into account. The municipal allocation recovers its aurora of attractiveness and has greater price resiliency, while the allocation for structured securities declines as rising rates lead to an extended duration and increased price depreciation. The decline in the high-yield allocation reflects prospective relative value. The continued prominence of the equities and alternatives category is due to (floating rate) bank loans.
SUMMARY AND CONCLUDING REMARKS
There are several caveats and key take-aways from the findings in Table III. First, the results are very data specific, particularly for the embedded yield-term structure of taxable and tax-exempt holdings. Second, the prospective term structure of taxable and tax-exempt opportunities is a key component when assessing capital market prospects relative to current holdings. Those term structures will be impacted by eventual tax rates and regulations, as well as macro-economic and political events. Finally, individual company decisions will depend upon their return metrics: after-tax book incomes versus after-tax total return.6
With a new sheriff in town, we caution about taking precipitous action or retreating from well-crafted strategies until the uncertainty abates. NEAM is ready to assist in your evaluation process - today and in the future. Look for an upcoming NEAM Quick Takes that focuses on macro-economic and political events and offers guidance on updating prospective returns. We hope these insights will be of use in helping you achieve the best outcomes for your organization.
1 See, “Considering Opportunities in Low Return Uncertain Environment – An Enterprise View,” Perspectives, September 2016.
2 In fixed income, all else equal (credit/defaults, maturities, optionality, etc.) it is all about the coupon.
3 Industry holdings’ cusips are assigned to one of 58 fixed income and several equity indices for which NEAM has constituents’ daily income, price and markets statistics since 1997. For fixed income, indices are categorized by sector and within sector by rating cohorts and duration tenors.
4 The estimated industry return follows the DuPont framework: Return on equity equals insurance product margin times the product leverage plus the investment return times the investment leverage, where the product margin and asset return are tax affected. In the example: 7.82 = 3 * (1-.35) *.76 + 2.85 (ROA shown after-tax) *2.2.
5 The average credit rating of the industry’s taxable bond portfolio is single A.
6 Whether the return objective is total return or book income, the T-VaR estimates we show are marked-to-market total return based. Accordingly, the impact of fixed income prospective returns might be ignored in the decision framework, relying on book income, assuming the real world impact of price volatility upon enterprise risk remains.