Estate Tax Planning After ATRA: What’s Left?

Joseph Septimus and Tara Thompson Popernik
Published Date:
Apr 1, 2015

For the overwhelming majority of US taxpayers, the American Taxpayer Relief Act of 2012 (“ATRA”) eliminated the need for federal estate tax planning. It did so by making the $5 million applicable exemption amount (indexed for inflation) permanent. Additionally, in April 2014, the New York State estate tax exemption increased, so fewer New York taxpayers would need state estate tax planning.

The benefit of inflation indexing cannot be understated. For example, since the enactment of ATRA, the applicable exemption amount has increased to $5,430,000 (2015), a $430,000 increase in just a few years. Over the next 25 years, in a typical inflationary environment, we project that the exemption will increase by an additional $5.8 million (Exhibit 1). Based on these projections, a couple currently in their early sixties may have an applicable exemption amount in excess of $22.4 million at the time they reach their mid-eighties. If inflation is higher, the exemption could be much higher.

Indexing the applicable exemption will eliminate all federal estate tax for the majority of the US population. In spite of not being subject to estate tax, heirs receive an additional income tax benefit: a step up to fair market value of the cost basis in the assets they inherit, as of the date of death. As a result, pre-death appreciation (the difference between the decedent’s adjusted cost basis and fair market value on the date of death) will never be taxed. There will only be taxable gain if there is appreciation after death, and then only to the extent of the post-death appreciation.

As long as this system remains in place, most wealthy families should focus on income tax, not estate tax, planning. While President Obama’s recent budget proposal includes eliminating the step-up in cost basis, it is unlikely this provision will be adopted.

As discussed below, there are two groups of taxpayers who would benefit from an estate plan aimed at reducing estate taxes and income tax for their heirs.  

Who Needs Estate Planning?

The two groups of taxpayers who will benefit from estate planning include (1) individuals with a current estate value in excess of the applicable exemption amount who are unlikely to consume enough of their wealth during their lifetime to avoid an estate tax problem, and (2) individuals who find their previous estate plans suboptimal with regard to income tax or a step-up in cost basis as a result of tax law changes.

This article addresses the concerns of both groups by focusing on an advanced planning technique that could significantly reduce a taxable estate over time: zeroed-out Grantor Retained Annuity Trusts (“GRATs”). It quantifies the potential benefit, using our Wealth Forecasting System.*

The goal of the strategy is to pass a remainder interest in a trust from the grantor to his or her intended beneficiaries at the lowest possible gift tax value. In doing so, the grantor retains an annual payment interest for a term of years from the trust, based on the minimum rate required by the Internal Revenue Code and applicable regulations. If at the end of the term, the actual value of the gift exceeds the gift tax value (established at the time the trust was created), the additional investment value would be transferred to the beneficiaries, estate and gift tax-free.

Caution: Before recommending this technique, make sure that your client is likely to have a taxable estate. As noted above, even in a very low-inflation environment, the applicable exemption amount is likely to increase to about $11.2 million per person over the next 25 years. Thus, if the client implements these strategies but is not subject to estate tax (because their estate is below the applicable exemption amount), the strategy would have accomplished nothing.  Even worse, the heirs who receive the gift of the remainder interest during the life of the grantor rather than at death (had the strategy not been implemented) would not be entitled to a step-up in cost basis to the fair value at the date of death.

How Zeroed-Out GRATs Work

In the current low interest rate environment, zeroed-out GRATs can provide significant gift and estate tax savings. For example, we recently helped a tax advisor forecast his client’s estate-tax liability 20 years from now as being about $22 million without additional planning—and half as much, or $11 million, by implementing a particular, “rolling” version of the zeroed-out GRAT strategy.

Most zeroed-out GRAT strategies have 10-year terms and entail four steps:

  • The grantor creates a trust (the GRAT) and contributes assets to the trust.
  • The grantor receives an annuity payment for each year of the GRAT term equal to an amortized principal payment plus the minimum rate of return as computed pursuant to Section 7520 of the Code. If at the time the GRAT is created, the amount contributed to the GRAT equals the grantor’s retained interest (i.e., the present value of the future annuity payments due to the grantor) the GRAT is “zeroed-out.” In other words, because the grantor’s retained interest over the term of years is equal to the remainder interest (the gift to the remainder beneficiaries), the remainder interest would be used up by the end of the term. Based on those assumptions, there would be no taxable gift and, thus, no gift tax liability.
  • During the GRAT term, the grantor receives the required annuity payments and pays income tax on the GRAT’s earnings.
  • At the end of the term, funds remaining in the GRAT are distributed to the beneficiaries or to a trust set up for the beneficiaries, passing from the trust free of gift and estate tax. If the beneficiaries are the grantor’s grandchildren, generation-skipping tax issues may apply.

But these GRATs pose significant capital-markets risk: Unless the investment return on the trust assets outperforms the Section 7520 rate (determined at the time the trust is created) over the term of the trust, there are no tax-free funds to be distributed to the beneficiaries. The Section 7520 rate is based on a formula that reflects current Treasury yields. For March 2015, the minimum rate is 1.8%, near its all-time low.

Although our research suggests stock market returns are highly likely to exceed 1.8% per year over the next 10 years, they could be lower. Because stock market returns are notoriously volatile, one or two years of extremely bad performance of the trust portfolio could cause the GRAT investment return to underperform the 7520 rate over the term of the trust. If this occurs, the trust corpus would run out with nothing for the remainder beneficiaries to receive at the end of the term.

For example, consider a 10-year GRAT established in early January 1999 with its corpus invested in US stocks (during the worst 10-year period of S&P 500 returns). In spite of positive investment performance in 6 of the 10 years (Exhibit 2), the beneficiaries would have received nothing at the end of the term, due to overall underperformance.

How Rolling GRATS Work

To reduce the risk of such a dire result, consider setting up a series of two-year rolling GRATs over a 10-year period, instead of one 10-year GRAT. Although the typical term of a GRAT is 10 years, the minimum term is 2 years.

In creating a series of rolling GRATs, the grantor contributes the initial assets to a trust with a two-year term and two annuity payouts: one at the end of year one; and the other at the end of year two. In turn, the grantor contributes some or all of each annuity payout to a new two-year GRAT. The grantor can choose to keep some or all of each annuity payment along the way. If desired, the grantor can terminate the strategy in as little as two years by ceasing to contribute annuity payments to a new GRAT.

By implementing this strategy, years of extreme negative performance can be isolated so that at least a few GRATs in the series will outperform the Section 7520 rate, even if the cumulative return over the 10-year period would have underperformed the initial 7520 rate. Exhibit 3 (covering the same 10-year period as Display 2) illustrates the overall benefits of creating two-year rolling GRATs to transfer wealth as compared to a single 10-year GRAT that would have left no remainder to the beneficiaries.

In a low interest environment, a common objection to using a rolling GRAT is that it does not lock in a low Section 7520 rate that drives down the gift tax value of the remainder interest.  Because each new GRAT is subject to the prevailing Section 7520 rate at the time it is established, it is possible that the assets of some or all of the subsequently created GRATs could have a higher interest rate hurdle to overcome. On the other hand, in the current, extremely low interest-rate environment, rising interest rates would likely reflect expectations of stronger economic growth ahead and potentially propel strong stock market returns. In our analysis, strong equity returns are likely to outperform the corresponding Section 7520 rates that may be a percent or two higher than the current rate.

In the end, the potential success or failure of this strategy depends on capital-markets returns. In working with clients and their tax advisors, we use our Wealth Forecasting System to forecast 10,000 potential scenarios to generate a range of likely outcomes.

Based on our forecast, a 10-year GRAT or a series of two-year rolling GRATs (over the same 10-year period) funded with a globally diversified portfolio of stocks are both highly likely to succeed in transferring significant wealth (Exhibit 4). However, the two-year rolling GRAT strategy is likely to transfer more than twice the amount of wealth in the median case; it is even more likely to succeed in adverse markets. We estimate the likelihood of passing on tax-free wealth to the next generation are 98% for the rolling GRAT, vs. 82% for the 10-year GRAT.  

*The Bernstein Wealth Forecasting System uses a Monte Carlo model that simulates 10,000 plausible paths of return for each asset class and inflation and produces a probability distribution of outcomes. The model does not draw randomly from a set of historical returns to produce estimates for the future. Instead, the forecasts (1) are based on the building blocks of asset returns, such as inflation, yields, yield spreads, stock earnings, and price multiples; (2) incorporate the linkages that exist among the returns of various asset classes; (3) take into account current market conditions at the beginning of the analysis; and (4) factor in a reasonable degree of randomness and unpredictability.

Bernstein does not provide tax, legal, or accounting advice. In considering this material, individuals should discuss their circumstances with professionals in those areas before making any decisions.

SeptimusJoseph Septimus is an Advisor at Bernstein, the Co-Chair of the Trusts & Estates Subcommittee of the New York State Bar Association’s Tax Section, and an adjunct professor of law at Brooklyn Law School. He can be reached at

ThompsonTara Thompson Popernik
is Research Director of the Wealth Planning and Analysis Group at Bernstein. She can be reached at

Views expressed in articles published in Tax Stringer are the authors' only and are not to be attributed to the publication, its editors, the NYSSCPA or FAE, or their directors, officers, or employees, unless expressly so stated. Articles contain information believed by the authors to be accurate, but the publisher, editors and authors are not engaged in redering legal, accounting or other professional services. If specific professional advice or assistance is required, the services of a competent professional should be sought.