Description

Summary Thesis:

I believe the mortgage insurance industry, and NMIH in particular is an attractive opportunity today.  Deep scars from the last housing driven recession and resulting PMI industry carnage are likely creating this opportunity.  Pre-COVID, the MI’s were writing high quality policies backed by extensive reinsurance.  Housing’s current trends and generous forbearance programs are likely to show that the worst of the provisioning has already occurred.  I think NMIH is at $40 by year end 2022, producing an attractive IRR, with likely higher than average volatility.

Definitions

• IIF - insurance in force - unpaid principal of mortgages insured by NMIH

• NIW- New Insurance Written - new unpaid principal originated in the period

• RIF- Risk In Force - IIF multiplied by the amount of coverage provided by NMIH.  This is the ultimate gross (pre reinsurance) loss exposure should all covered loans default

• PMIERs - Private mortgage insurer eligibility requirements are a set of guidelines issued by the GSEs for PMI providers.  The most discussed component is required assets, which PMI insurers are required to keep liquid funds available in excess of this amount.  Required assets is a risk adjusted metric based on the quality of the insured portfolio and reinsurance coverage

Brief Industry Description:

Most are likely familiar with mortgage insurance (private mortgage insurance - PMI).  As spelled out in FNMA / FMCC charter’s, the GSE’s are required to obtain credit enhancement for any loan made that exceeds an 80% LTV.  Credit enhancement options include obtaining insurance from a GSE approved insurer (PMI), the originator retaining 10% of the loan, or an agreement by the mortgage seller to repurchase or replace the loan upon default.  As likely expected, most originators choose PMI.

PMI is not the only option for homeowners making a low down payment purchase, the federal government is also an active insurer through the FHA, USDA, and VA.    Borrowers with lower credit scores and/or higher LTVs are likely to find FHA insurance cheaper than PMI.  This skews PMI users to the higher credit quality, lower LTV cohorts of low down payment borrowers.  

Pre-financial crisis PMI was the predominant source of insurance.  In 2007, the industry wrote $357 billion in insurance (which includes the total notional amount of the loans insured).  However, lax underwriting standards by most participants pre-crisis decimated the industry, wiping out capital, PMI competitors and the goodwill built with the GSE’s.  Federal insurance filled the void as PMI players stepped back.  But this share has been shifting back again, with PMI gaining share (presently about equally split between private and federal programs).  The image below provides added context for how badly the GFC hurt the industry.

The industry has shifted dramatically since the crisis.  Three of the eight firms writing business were taken over by state regulators, while Essent, Arch and National MI entered the scene either through acquiring legacy platforms or building from the ground up.  Today there are six players, all evenly split at around 17-19% market share, except for NMIH which is at 10%.  The makeup of the business also shifted post crisis to higher FICO’s, lower LTV, and lower DTI.  For context, 38% of insurance in force FICO’s in 2007 were under 680 and 33% of insurance in force LTVs were above 95%.  The chart below shows new insurance written in 2019 for these metrics.

In addition to a better mix of new business, the industry has also been utilizing extensive reinsurance, including quota share, excess of loss and insurance linked notes. For the sake of brevity, I will not go into detail here, happy to elaborate in comment section.

Brief Business Description: 

NMIH is a holding company with three primary subsidiaries: National Mortgage Insurance Corporation (National MI), National Mortgage Reinsurance, and NMI Services, a small business focused on outsourced loan review.  NMIH was incorporated in 2011 and wrote its first policy in 2013.  NMIH’s chairman and former CEO, Brad Shuster, and current CEO Claudia Merkle were executives at PMI Group, a legacy MI writer who filed for bankruptcy in 2011.  Major stockholders at the time of the 2013 IPO included Blue Mountain Capital, Claren Road, Hayman Capital, Perry, Amici and Columbus Hill.  

The following graphics show how NMIH is positioned relative to peers.

As can be seen with the portfolio characteristics from earlier and the risk assets to risk in force ratio (with the numerator adjusted for reinsurance and portfolio attributes) NMIH has a strong position.

Key Positive Points:

This crisis is very different from the last and I expect the outcome to be very different for the PMI players.  The crux of the bullish argument rests on the continued improvement in severity and frequency trends seen in NMIH’s portfolio driven by home price stability, continued repair of consumer finances and forbearance programs.

Home Prices

Home prices are a key ingredient to any PMI insurer - as this directly impacts the consumers decision to default.  A homeowner with equity in the home would rather sell the home if possible, pocket their equity and salvage their credit.  Mortgage insurers are expecting 100% severity on future claims.  The importance of this variable should be considered when looking at the chart above displaying industry profitability.  The industry has, in general, been profitable most of the time, despite the period shown containing five recessions.  We have seen home prices accelerate this recession to date.  Looking at the intermediate term it is possible we see moderation in this trend.  I agree with most that we aren’t likely to see employment reach pre-pandemic levels soon, which potentially removes bidders and adds sellers to the single family market.  When contemplating the future most PMI industry participants are using moderate home price declines as a base case:

NMIH CFO Q3 2020 Call - “We continue to assume that house prices will be down modestly over the next 2 years, we think it's an appropriately conservative stance.”

ESNT CEO Q2 2020 Call - ”When we entered into the pandemic, our view is home prices would flatten and start to decline, which is really still our view.”

ACGL CFO September 2020 Barclays Conference - “And that's always something we've said from the get-go is home prices are, by far, the most important variable to the performance of the book.”

The following slide shows the importance of home prices for severity..  The slide was displayed at their 11/19/2020 investor day:

Consumer Balance Sheet / Forbearance

The CAREs Act put in place a forbearance program for homeowners experiencing financial hardship due to the coronavirus pandemic.  Homeowners requesting forbearance can forgo paying principal and interest for 180 days, and have the option for another 180 day extension.  At the end of the period homeowners have three main options to get current again: 

1. Reinstatement-pay any missed payments at the end of the forbearance period

2. Repayment – pay the missed payments off over time by adding to the monthly payment

3. COVID 19 Payment Deferral – add the missed amount to the end of the mortgage term

An important point when considering NMIH is that mortgage insurers book provisions when a covered loan misses two payments.  Therefore, loans in forbearance that aren’t paying are included in NMIH defaults and are getting reserved for.

The chart below shows forbearance and default activity for NMIH.  Defaults peaked in August at 14,236.  As of October 31st, total defaults stood at 13,108.

An important aspect of this table is the number of loans in forbearance that have missed one payment. This is essentially the pipeline for future defaults, as most defaults are in forbearance.  Forbearance loans in default will likely continue to be the vast majority of all defaults until the application window expires, currently expected on 12/31/2020.

Here are some notes from Black Knight:

• Rolls from current to 30-day delinquent is now below pre-pandemic levels, suggesting continued improvement in new defaults

• 6.3 million borrowers entered COVID-19 related forbearance, ½ have since exited

• Of the 3 million borrowers remaining in forbearance, 2.4 million have extended, the remainder are in their first program term.  GSE’s represent approximately 1 million of this total.

• Of the loans in forbearance that are guaranteed by GSE’s, only 4% are loans with less than 10% equity.

For NMIH, I expect quarterly defaults to move down from the 6,588 new defaults recognized in Q3 to close to 2,000 in Q4 and then trending lower back to a normal level maybe in 2023.  In summary, forbearance did a great job of carrying most through the extreme shock witnessed in April, allowing most to regain their footing.  But - without 100% of laid off folks regaining employment, there will likely be elevated defaults relative to pre-COVID levels for a few years.  See the section below for future fundamental outlook.

Other Potential Positives:

There are two other industry positives, and one NMIH specific positive that I think are worthy of mention.

Book of Business

Given the decline in rates, the book is seeing significant churn.  As of the most recent quarter, persistence (percentage of insurance in force that remains after 12-months) stood at 60.0%, down from levels around the 80’s a few years back.  This is creating churn in the book and lowering the average life of a policy.  However, insurance in force has grown given the strong purchase and refinance mortgage markets.  As rates move higher, the persistency will again expand and lengthen the life of a policy. 

This portfolio churn can be coupled with improved pricing and higher quality risk.  Insured loans today have a higher quality compared to those written pre-pandemic.  NMIH’s weighted average NIW FICO increased to 764 in Q3  from 757 in Q1, weighted average LTV is at 90.7% in Q3, down from 91.3% in Q1, and debt to income was 32.6% in Q3 down from 34.4% in Q1.  Pricing was also boosted across the industry, with prices increasing 10% plus.

Down Payment Affordability 

Down payments are a significant hurdle for many new homeowners.  According to the FHA, the percentage of borrowers receiving assistance (often family) with the downpayment is growing.  I expect this trend to continue and likely provide a tailwind to PMI players.

Is Arch Pulling Back?

Arch is the industry leader, however, it also has a sizable insurance / reinsurance book (see tdylan409 writeup from May).  Though I don’t expect a massive shift in posture from ACGL, but multiple comments on the last call implied improved opportunities outside of MI.  All else equal capital shifted to support reinsurance and insurance may open an opportunity for other PMI players to grab share.

Key Areas of Concern:

Regulatory

The major concern for me is regulatory.  

A long stated and well-intended goal of having more private capital at risk in the U.S. housing mortgage market creates the possibility of future turbulence for the MI industry.  PMI is obviously a source of private risk capital, but, given how the industry performed during the GFC they don’t have the best reputation.  However, PMI still is a large portion of private risk transfer capital (from Q2 FNMA presentation):

The need for PMI is written into the GSE charters, so more back-end CRT through CAS and STACR is not a major concern.  However, GSE programs like Imagin and EPMI do directly encroach on MI territory.

Under these programs, which are run by Arch, as an alternative to securing MI, lenders can ship high LTV loans to Freddie/Fannie without PMI.   The risk is 100% transferred to reinsurers.  Arch stated that this business was around 2% of market NIW in their Q4 2018 earnings call.  This is a program to watch. 

PMI insurers really upset the industry during the last downturn, if they botch the current episode it could be game over.

Fundamentals & Price Outlook:

Starting with the industry mortgage originations.  I have no differentiated viewpoint here, and I think MBA Forecasts are reasonable.  Originations are much higher this year, both for purchases, but especially so for refinancings (2020 estimates +15.8%, and +70.6%, respectively).  Going forward, it's a near certainty this will step back.

Mortgage insurance as a proportion of total originations has been rising fairly dramatically in recent years, from 13.2% in 2016 to 18.6% last year (per ESNT 10-k).  Share gains are likely again this year with industry NIW in excess of $550 billion v. $384b last year.

The following is my base case outlook:

NMIH figures are in millions, industry figures are in billions.

From a pricing perspective, there are a couple of competing dynamics. New policy pricing is heading higher as explained earlier, however offsetting this is pressure from Tax Cuts and Job Act pricing and gains from single policy cancelation.  Due to the reduction in corporate tax rates, the industry repriced lower.  Since the pricing doesn’t change for existing policies, the fact that pre TCJA policies are still on the books creates a headwind as those roll off.  As an example, ESNT’s 2016 premium yield was in the mid 50 bps range, and it has recently been in the mid 40’s. 

Another headwind has been the large windfall received from single premium policy lapses.  Most policies are paid monthly, but some are up-front.  When an up front policy cancels prematurely (refinanced), all future revenue is recognized in that period.  Given the level of refinancings, this has been a boost to premium yields.  NMIH reported a net premium yield of 0.39% annualized in the latest quarter, absent the $12.6 million in single premium cancellation, the yield would have dropped another 5 bps  to 0.34%.

The last component impacting yields is reinsurance, how this influences future yield is management's decision and I don’t expect major changes from this variable.

In summary, I expect pricing to move lower by a basis point or two as the single cancellation benefit likely offsets the new pricing.

The last of the most important fundamental variables is defaults, losses and provisions.  Provisions can be triangulated by contemplating the default rate, the percentage of defaults that become actual claims, and severity.  Severity is assumed to be 100%.  As stated above, I expect new defaults to taper meaningfully starting in the fourth quarter and to decline >60% next year (though still up 2.7x from 2019).  Reserve per default was $16,000 last year, and is running at $6,000 this year, given COVID defaults lower propensity to turn into a claim.  Provision per new default this quarter was $2,800 compared to $5,000 last year.  I expect reserves per default to dip in 2020, rebound in 2021, and resume a home price like trajectory moving forward.

Looking at my estimates versus JPM it appears the biggest discrepancy is claims expense.  I would love to get others' perspectives.  Goldman has ACGL’s loss expense declining 30% in 2021, while JPM is looking for NMIH’s to grow 60% to $120 million. 

As displayed, I belief a 2023 EPS figure greater than $4.70 is achievable.  To help frame a price target, RDN, MTG, and ESNT have traded around 8.4x, while ESNT which lacks legacy exposure has traded around 9.2x.  These multiple suggest a 2022 YE price around $40.

As an added bonus - Without going into a ton of detail, NMIH is likely to pay a much lower cash tax than the 20% assumed above.  MI’s are required to place 50% of net earned premium in statutory contingency reserves, and they are able to deduct the amount contributed if they put the tax benefit received in tax and loss bonds.  The reserves are required to be held for 10 years, unless incurred losses exceed 35% of net premiums.  Insurers then pay the tax once the reserves are released.  With a growing book of business, this effectively punts a large portion of taxes into the future.

Catalyst

• No unique catalyst