艾芙隆海灣社區公司 (AVB) 2017 Q3 法說會逐字稿

Good morning, and welcome to the AvalonBay Communities Third Quarter 2017 Earnings Conference Call. (Operator Instructions) Your host for today's conference call is Mr. Jason Reilley, Senior Director of Investor Relations. Mr. Reilley, you may begin your conference.

Thank you, Aaron, and welcome to AvalonBay Communities Third Quarter 2017 Earnings Conference Call. Before we begin, please note that forward-looking statements may be made during this discussion. There are a variety of risks and uncertainties associated with forward-looking statements, and actual results may differ materially. There's a discussion of these risks and uncertainties in yesterday afternoon's press release as well as in the company's Form 10-K and Form 10-Q filed with the SEC. As usual, this press release does include an attachment with definitions and reconciliations of the non-GAAP financial measures and other terms which may be used in today's discussion. The attachment is also available on our website at www.avalonbay.com/earnings, and we encourage you to refer to this information during the review of our operating results and financial performance. And with that, I'll turn the call over to Tim Naughton, Chairman and CEO of AvalonBay Communities, for his remarks. Tim?

All right. Thanks, Jason, and welcome to our Q3 call. With me today are Kevin O'Shea, Sean Breslin and Matt Birenbaum. Each of us will provide some comments on the slides that we posted last night, and then we'll all be available for Q&A afterwards.

I'll start with a brief overview of Q3 results and then Sean will provide an update on operations and how the portfolio is positioned headed into Q4. Matt will talk about development activity and how we are positioned to deliver growth through that platform over the next 2 to 3 years, and Kevin will discuss the closeout of Fund II and provide some insight into the balance sheet at this point of the cycle. And lastly, I'll wrap up with some thoughts about our recent announcement to enter the Denver market and our intention to enter Southeast Florida in the future.

We're just starting now on Slide 4. Results for the quarter include that core FFO per share growth of 5.8%. Same-store revenue growth was at 2.2% or 2.3% including redevelopment. Year-to-date, same-store revenue growth stands at 2.6% or 2.8% including redevelopment. And then lastly, same-store sequential growth came in at 1.3% for the quarter and 1.4% once you include redevelopment.

We completed another $95 million of new development in Q3, bringing total completed year-to-date to $1.15 billion. We expect to complete another $650 million or so in Q4, which will result in about $1.8 billion completed for the full year.

Year-to-date and full year completions are expected to stabilize at an initial yield at a low 6% range, well above prevailing cap rates for this basket of assets. In his comments, Matt will describe in more detail how these completions, along with development under way, will drive earnings and NAV growth over the next couple of years.

I'll now turn it over to Sean who'll discuss portfolio operating results.

All right. Thanks, Tim. Turning to Slide 5. Our same-store portfolio continues to produce rent change that's generally consistent with supply and demand being essentially in equilibrium. We've been trending in the 2% to 2.5% range all year long with renewals consistently in the low 4s and new move-ins following a more seasonal pattern. For Q3, renewals averaged 4.1% while new move-ins were 90 basis points.

In terms of the regions, Seattle continued to lead the way with rent change north of 5% for the third quarter, followed by Southern California in the low 4% range. Boston produced rent change in the low 3s, while Northern California, the Mid-Atlantic and the Greater New York, New Jersey regions all delivered rent change in the 1.5% to 2% range.

Moving to Slide 6. The same-store portfolio is very well positioned for the seasonally slower fourth quarter. Our same-store average physical occupancy rate for October is about 20 basis points above last year and availability is trending roughly 30 basis points lower.

For the development portfolio, we met our revenue and NOI expectations for the third quarter. In addition, our deliveries during the third quarter were consistent with our midyear update. So we're comfortable with the expected pace of both deliveries and absorption for the fourth quarter.

And with that, I'll turn it over to Matt to talk about development in more detail.

Great. Thanks, Sean. Turning to development. Our development underway should provide good future growth to both earnings and NAV as those communities are completed and finished their lease-ups.

Slide 7 shows the $225 million in NOI that we're expecting to realize from our current development, with nearly half of that NOI to come from communities that have not even started any leasing activity. Only $17 million worth of NOI was generated by these properties in the current quarter but there's clearly a lot of additional NOI to come.

On Slide 8, we translate this future NOI into core FFO. We've raised most, but not all, of the capital required to fund this $3.8 billion book of business. Because the earnings accretion depends on what the ultimate cost of that capital turns out to be, we've shown a range of earnings impacts here based on weighted average initial cost of capital between 3.5% and 4%, which in turn translates into $0.53 to $0.67 of incremental core FFO upon stabilization.

Slide 9 shows the spot value creation for this development translated into NAV per share accretion, the current weighted average initial yield on this basket of properties is 5.9% and we believe if these assets were completed and available for sale in today's market, they would sell for a weighted average cap rate of roughly 4.4%. This translates into $1.3 billion of value creation on completion or $9.42 per share.

Turning to Slide 10. We have a very long track record of delivering strong risk-adjusted returns from our development activity through all stages of multiple market cycles. As Tim mentioned, we expect to deliver a company record of $1.8 billion in development completions in 2017 at profit margins of close to 30%, consistent with the margins we have achieved throughout this cycle.

The chart on the left shows our development yield by year of completion going all the way back to the late 1990s and compares those yields to cap rates in that same year. With the exception of the single year of 2009 at the height of the global financial crisis, we have delivered a positive spot value creation on completion every year for 20 years running. And because we're a long-term investors and not merchant builders, even those deals completed in the worst-case scenario year of 2009 are today yielding 7% on cost, allowing for plenty of long-term value creation from that book of business as well.

The long-term returns can be seen on the chart on the right where we have sorted our unlevered development IRRs by the point in various cycles at which they were completed. While the deals which were completed earlier in the cycle have delivered stronger long-term returns, the returns across all time periods have been well in excess of our cost of capital and have delivered strong NAV creation for our shareholders.

Now I'll turn it over to Kevin to discuss Fund II performance and the balance sheet.

Thanks, Matt. Matt just touched on the primary way in which we allocate capital to new investments, which is through development. Another way we do so is through acquisitions. In this cycle, one vehicle we've deployed to create value through acquisitions is our second value-added fund, for which we commenced investment activity in 2009 and sold our last community this past quarter.

In addition to receiving asset management and property management fees, AvalonBay, as a general partner, received attractive investment returns as well as a promoted return of $35 million above our share of the fund's investment returns for performance achieved in excess of certain thresholds.

As you can see on the bottom of Slide 11, AvalonBay's investment returns on Fund II were strong. Excluding fee income, AvalonBay achieved a gross levered cash flow multiple on its invested equity of 2.4x and a gross levered internal rate of return of 19.2%.

Turning briefly to the balance sheet on Slide 12. We thought it might be helpful to highlight the evolution of a few balance sheet metrics across the cycle. Specifically, we show how 4 items, our net debt-to-EBITDA ratio, our unencumbered NOI percentage, composition of our debt and our credit rating have evolved over 3 periods; 4Q 2009, during the last downturn, 4Q 2013 not long after we completed the Archstone acquisition; and then today. As you can see, our credit profile and financial flexibility have significantly improved over the cycle and are arguably as strong as they have ever been in the company's history, which bodes well for our capacity to provide continued strength, stability and growth throughout the cycle.

And now I'll turn it back to Tim to discuss our market expansion into Denver and southeast Florida.

Thanks, Kevin, and we're now on Slide 13. As you know, we previously announced our entry into the Denver market with a recent acquisition. In addition, today we're announcing our intention to expand into Southeast Florida as well. We've talked with a number of you over the years about our geographic footprint, and some of the key factors that inform our thoughts with respect to market selection and portfolio allocation. We thought it would be helpful to provide a little more color behind our decision to expand into these 2 markets.

First, I guess I'd like to preface my comments by saying that while we are generally viewed as having had a stable strategy over the years, it has not been a static one. In fact, over time, our strategy has evolved to expand into high-rise and mixed-use products, which allowed us to enter many urban TOD and infill suburban submarkets. And through our brands, we now target multiple customer segments including a value-oriented customer through the eaves brand and a younger social seeking demographic through AVA. And we have entered and exited markets over the years, entering markets like Seattle and Baltimore while exiting markets like Chicago and Minneapolis.

So our decision to enter Denver and Southeast Florida does not reflect a shift in strategy but rather an evolution and implementation of our strategy. And when it comes to market selection, some of the key attributes that we look for and demonstrate on this slide include markets with, first and foremost, a higher percentage of knowledge-based employment, over-indexing and [STEM], finance, education and health-care-related jobs. Increasingly, we believe that there will be winning and losing MSAs as the economy migrates to more of a knowledge-based economy.

A second and related factor is that we favor markets that appeal to our target customer profile, as we prefer markets that again over-index to the younger college-educated and higher income boomer segments.

Third, we prefer markets that are characterized by lower housing affordability which helps support demand for rental housing and higher rental propensities.

Fourth, we prefer markets where the public sector is active and invest in infrastructure and cultural amenities, both important factors to support growth and quality of life within the market.

And lastly, importantly, we like markets that we believe play to our competitive advantages as a value-add investor, markets characterized by tougher and then lengthier entitlement processes that help to constrain or regulate supply, where a talented developer can grow its portfolio and create significant value through a deep level of understanding of its markets and skillful navigation of local politics.

Moving now to Slide 14. Of course, these attributes hopefully are correlated with markets that demonstrate strong rent growth over the long term, which is a critical component of return for our long-term investor and something we consider to validate these preferred attributes. As you see from this slide, over the last 15 years, Denver and Southeast Florida have actually compared favorably to our overall portfolio. This chart reflects market level rent growth over that time. Performance at the REIT level has been somewhat stronger. As you can see you, Southeast Florida has performed stronger than our East Coast footprint, while Denver has performed in line with our Western markets.

Now to Slide 15. One last point I'd like to make is that these expansion markets enhance overall portfolio diversification. Southeast Florida, in particular, is not well correlated with our existing markets and both markets provide some diversification from the areas of greatest concentration, that being California and the northeast. At target allocation levels, the portfolio is roughly split into 3/3: 1/3 California, 1/3 Northeast and 1/3 higher growth markets.

In terms of market penetration strategy, we expect that we'll invest both through acquisitions and development and, in addition, may choose to capitalize local developers to initially leverage their market knowledge and pipeline to help accelerate our expansion objectives. Of course, our tactics will ultimately be a function of value and opportunity as we have no set time table to reach any target allocation. Smart capital allocation will remain the priority. It's always been and we'll look to deploy capital appropriately across the cycle.

So in summary, turning to Slide 16, it was a solid quarter with core FFO above our outlook and the portfolio is well positioned headed into Q4. Development deliveries are tracking our midyear re-forecast and development should contribute meaningfully to FFO and NAV growth over the next 2 to 3 years. Our balance sheet is well positioned to fund additional growth and protect the company from downside economic scenarios. And lastly, our strategy continues to evolve as we're excited to add Denver and Southeast Florida to our market footprint as we believe those markets contain many of the key attributes that allow us to create even more value for shareholders over time. So with that, Aaron, we'd be happy to open the call for Q&A

(Operator Instructions) And we'll go first to Nick Joseph with Citi.

I appreciate your comments on the market expansion but just wondering why now is the right time to expand to these new markets? And how did you weigh market expansion against going deeper into your existing markets?

Nick, yes. A fair question from a timing standpoint. Obviously, we're several years into the current cycle. Denver, in particular, has had a very nice run this cycle. I guess I'd say a couple of things. One, as it relates to acquisition -- acquisitions, we will be more cautious with respect -- in particular with respect to submarket focus. We're really trying to focus more in the markets that we think are bit more supply constrained given the sort of current recent trends. As it relates to development, as I mentioned, we do anticipate using really both acquisitions and development. I guess I'd say these markets, as well as our existing footprint, the land frenzy is sort of behind us at this point. And we've also learned over time that to take advantage of any distress or downturn that might occur from any economic correction, you need to be in business. And so it's our intention to establish sort of beachheads in both those regions so that we are positioned to take advantage when the markets do turn. And then I did mention willingness to look at development JVs, which we do think is a good way to tap into deals maybe that might deliver 2 to 3 years out. And so when you think of sort of the mix of time frames that you might get from acquisitions, development JVs or our own sponsored development, I think -- we think it sort of diversifies kind of our risk against any cycle risk that may be out there as they have time frames from now to 4 or 5 years from now. Our long-term objective is to have a full-service office in each region. And I think one of our key competencies has been able to build great teams in our regions and to support them appropriately essentially. As it relates to our existing markets, we're always looking to penetrate those markets more deeply. So it's always something to consider. And we don't think we're missing out on opportunities today in our current markets but we do think there's an opportunity to expand our horizons a bit. As I mentioned earlier, we favor knowledge-based economies, and AvalonBay's markets, while we're heavily indexed to that, we don't have a monopoly on knowledge-based jobs. I think it's kind of interesting. I think the Amazon HQ2 project is sort of evidence of that. Recently, I've heard that Google, who, as I understand, allows employees to move to almost to any market they want, shut down Denver recently because it was such a popular destination. So it's a recognition that both -- I think both these regions in a way provide a bit of a release valve to Northeast, in the case of Florida, Southeast Florida and Denver in the case of California and particularly Northern California. So I don't know if that's entirely responsive to your question, but that's just a few thoughts.

That was very helpful. And then just one follow-up. Were there any markets that you considered expanding into? And could further expansion be announced in the near or medium term?

Yes. Thanks, Nick. Thanks for the follow-up. Well, first of all, I mean every couple of years, we always take sort of a comprehensive look at our markets, and it's something we've been thinking about, honestly, for the last 10 or 15 years whether we have the right footprint. In most cases, it's -- it resulted in us maybe exiting markets as opposed to entering markets and doubling down on our existing footprint sort of along the lines, if that's really the question. But one of the reasons we are announcing Southeast Florida today is because we did get that question a lot after we announced the acquisition of Denver, and this represents the 2 markets that we made the decision on today. So I don't think you should expect that there's any near-term announcement of further expansion beyond this.

We'll go next to Juan Sanabria with Bank of America.

Latest thoughts on supply, kind of what we should expect for '18 versus deliveries in '17. There's been some slippage and which markets across your main regions stand out as improving or getting weaker in '18, as you see it presently?

Yes. Juan, this is Sean. Happy to address that. What I can tell you is based on what people are communicating to us at present regarding the delivery of communities that may have under construction is for deliveries in 2017 it's about 89,000 units. In our footprint, that's about 2.1% of inventory. And based on what's on the books today that we can see across our footprint, in 2018, it's about 98,000 units or about 2.3% of inventory. Based on recent history, particularly the difficult construction environment today in terms of labor availability and things of that sort and recent precedent in terms of delays that we're hearing from multiple developers out there, I'd expect the 2.1% for 2017 to actually come down a little bit as we get through the end of the year into January and look what actually completed, probably the high 1% range would be my guess is where it comes out. And so across the footprint we would expect an increase in supply next year. My guess is [2018] (corrected by company after the call) will look like 2.3% to 2.5% maybe when you get to January. But what actually gets delivered next year probably will be less than that. But if you just look across the footprint, probably more supply in '18 versus '17. As it relates to the regional distribution, I can tell you that New England is expected to come down about 60 basis points next year, primarily a function of supply in the Boston market coming down. And then all the other markets are flat to up a little bit. The ones that are -- the markets that are probably going to be up more significantly are the Pacific Northwest. It's projected to be up about 40 basis points from 3.5 to 3.9. And in Southern California across all 3 major markets there, L.A., Orange County, San Diego, it's expected to be up about 40 basis points. Most of that relates to an increase in supply in the Los Angeles market. So what I'd generally say is a little more supply in '18 with the exception of the markets that I mentioned in Boston in particular. And when you see a slowdown, it will really be second half of '18 deliveries before you start to see that. We're running about 60 basis points of inventory on a quarterly basis right now. That's not expected to fall off until you get to basically to Q3 of next year, so Q3 and Q4 start to fall off. So that's sort of the broad way to think about it, if that's helpful.

That is. And just to follow up to that, if you can comment on your expectations for Northern California. And then -- and also have you seen any impact on the availability of labor post the hurricanes in trying to complete developments?

Yes. Why don't I take the first one and we can talk about the second one as well, probably Matt or Tim. But first one, when you say expectations, are you talking about supply expectations?

Yes.

Okay. Yes. No problem. So in terms of supply expectations, if you look at '18, the only market that really of the 3 is expected to see a deceleration in deliveries is San Jose. It's running a little north of 3% of inventory right now. And for next year, we're expecting to be in the low 2% range. So we expect to see a pretty material drop-off in deliveries in the San Jose market in '18 but it's relatively flat when you look at the East Bay and when you look at San Francisco.

This is Matt. I can speak a little bit to the second question about labor availability and the impact of the hurricanes. I'd say it's probably too early to tell. It certainly seems reasonable to assume that it's going to impact the labor market, particularly for construction, particularly for less skilled construction labor. We haven't really seen that yet. Obviously, we're not building in the markets that were most directly impacted by the hurricane, but certainly, it's something to watch for as the rebuilding efforts get started, which usually there's a little bit of a lag effect on that.

We'll take the next question from Rich Hightower with Evercore ISI.

Sean, really quickly, if you could run through new and renewals by market in the third quarter, and I have a follow-up to that.

Yes. As opposed to going through each one of the markets. There's 16 markets out there across 6 regions. Happy to share that with you maybe offline as opposed to spending a lot of time on that. But broadly speaking, across the portfolio, rent change for the third quarter was 2.5%, which is 4 1 on renewals and 90 basis points on move-in, as I mentioned in my prepared remarks. But if that's okay, we can take the rest of the detail offline.

Yes. No, that's fine. I appreciate that. Second question, in the presentation, the $111 million of NOI from development not in lease-up, can you give us a sense of when over the course of the future that's supposed to hit, roughly?

I think that represents kind of all the deals that are on our development attachment that haven't yet started leasing. So if you go to that attachment 8, it will lay out kind of which quarter they'll start to expect leasing. But it's basically, over the next, call it, 2 years, they'll go from starting leasing to stabilization.

We'll go next to Nick Yulico with UBS.

In the past, you've talked about having around $1 billion of development starts annually. I'm wondering if that's still a good number and whether you're thinking about raising that level because of signs of the apartment cycle lasting longer than expected.

Nick, this is Tim. If you look at our development right pipeline, it kind of telegraphs kind of where development volumes are headed. I think we have about $3.2 billion in development pipeline that tends to be 3 years, 3-plus, maybe 3.5 years worth of pipeline. So it's still kind of in that $1 billion range, maybe a little less. As we said in the past, we think it's probably going to trail off a bit just given, even though the cycle is going longer. The economics are less compelling and less deals are sort of making it through the screen, if you will. And you're kind of hitting targets and we saw maybe earlier in the cycle just given [where] related construction costs have gone and are going right now relative to rent.

Okay. One other question is for the future development project you have on 96th Street and Second Avenue there's been some press about this that the governor is kind of looking into the whether it's park land or not and whether the project can go forward. Any sense on when and how this may get resolved? And if you can just remind us what the level of your investment is in that project? How much of that development right that you have on -- in the supplemental for New York accounts for that project?

Sure, Nick. This is Matt. Yes. That's an interesting question about what's the difference between a park and a playground, which apparently there's a legal distinction there, which matters to the governor and the mayor. So we fully expect that, that will get worked out but it will take some time. I don't think we have great visibility yet in terms of how much time that might add to the pre-development schedule. That is not a project that we expected to start in the next year, so we really view that as a next cycle deal in some ways. Maybe it's 2 to 3 years away from starting depending on how everything shakes out. It does represent, I want to say, it's about $700 million, $650 million of the $3.2 billion in development rights. And just to be clear, that is a ground lease, so that actually doesn't reflect the value of the real estate.

You take the next question from Austin Wurschmidt from KeyBanc Capital Markets.

Just looking at the like-term rent growth that you outlined in your slide deck. Tracking a little bit ahead of last year it looks like or in-line-ish. Was just curious what markets are really driving the inflection. And then with the supply backdrop that you talked about, do you think that, that remains stable in the next year? Or we could see a pullback a bit?

Yes. This is Sean. Just to be clear, as it relates to rent changes this year versus last year -- Jason could walk through the numbers with you but overall rent change is down this year relative to last year. It's a function of really 2 things. One is the supply that we talked about; and two, it's a slower pace of job creation generally across the U.S. and in our markets. So the rent change is down. In terms of the infection point, I think the point we're trying to make is that rent change appears to be leveling off. We're basically around 2% to 2.5% all year long. And we certainly aren't really talking much about guidance as it relates to next year but tried to provide some context on the supply side in terms of what we might expect, which is similar to essentially slighter greater supply in '18 as it relates to '17.

Maybe just to add. This is Tim. In terms of the macro -- we're not going to give guidance for next year certainly. But in terms of the macro environment, I guess we probably see more cross winds than anything else. It's our expectation, Sean mentioned earlier, supply to tick up a little bit. I think all things being equal, we probably expect job growth to tick down a little bit just given unemployment being at such a low rate right now and the difficulty of finding skilled labor. But that is offset by some sort of positive trends as well. The consumer and business confidence continues to increase. Labor participation rates are on the rise a little bit and wage growth is on the rise as well. We think all these things would help us stimulate household formation. So I think there's probably going to be some offsets, but in general, we think it's probably shaping up to be a decent environment for 2018 without getting into much more detail in terms of how it might be translating to our portfolio.

And then just lastly for me, I was just curious what your thoughts what you attribute the above average rent growth in both Southeast Florida and Denver over the last 15 years versus kind of the overall portfolio?

Well, really it's been on the demand side. The markets haven't suffered from the supply but Denver probably has more than it should have right now. The Sun Belt as actually performed quite well this cycle. Supply growth hasn't been, at least in the Sun Belt, hasn't been that much greater than some of our coastal markets, and it's had some pretty strong growth even in some of the areas that we thought our markets captured a lot of the higher tech jobs, we see a lot of that accumulate in the Sun Belt market and certainly that's true of Denver as well. I think they've become -- Denver's become somewhat like we saw Seattle 10 years ago, becoming another kind of tech anchor in the Western U.S., and Southeast Florida certainly has benefited from some of the anti-growth initiatives perhaps in the Northeast over the last cycle or so.

And then just one quick follow-up to that. I'm just curious how you think about the volatility in these markets through the cycle versus the existing portfolio.

Yes. I know. Great question. So Denver would tend to be more volatile and Southeast Florida would tend to be more stable. So as I mentioned in my remarks, Southeast Florida, like D.C., is the one market that's sort of negative or not well correlated with the rest of our footprint. So there's some appeal to us to that and then whereas Denver would be, would tend to track more what you see in the West Coast which means you've got to be careful as to how you allocate capital over the course of the cycle and be a lot more thoughtful from a timing perspective.

We'll take our next question from John Kim with BMO Capital Markets.

GGP discussed on its earnings call a JV agreement with you to build multi-family in one of their Seattle malls. How big of an opportunity is this for you as far as developing in existing retail?

Yes. This is Tim here. Yes. We are aware that GGP announced an agreement of principle on a single project in northern Seattle. Every time there's been a structural change in the economy it's resulted in massive repurposing of real estate, it's tended to happen over a longer period of time. And you move from agricultural to manufacturing as you move from manufacturer to more information-based economy. I think there are number of trends right now that we're going to -- that's just going to accelerate. Whether it's happening in the area of digital commerce, which is relevant to GGP, but in terms of things we're seeing in the sharing economy, the blurring of the work with play, just the nature of work changing just based upon knowledge-based jobs, I think that's -- we think that's going to have a big impact in terms of the need to repurpose real estate over the next 10 to 15 years. And a lot of great real estate has been consolidated over the last 15 or 20 years. And so we think there is an opportunity to partner with some of these owners that have consolidated great real estate during that period of time when high density housing is part of the solution when it comes to repurposing. In terms of the form of that partnership, it's going to take different forms, than it has already on deals that we've done. It might just mean helping them replan it and splitting the properties just take a piece of it for residential for ourselves and letting the existing owner sort of replan or take the rest of the property. It may be more of a condo structure like we had at Assembly Row with Federal where we may be building the retail and deeding it back to them sort of in a condo structure, where we own the residential, they own the retail. Or maybe a JV structure which we were working through with the GGP where we own together the residential component. And generally, we'll be amenable to that structure, where we think we have a partner on this, where there's a clear alignment of interest, where long-term owners both believe sort of in the long-term viability of that location in the real estate. But we think it's going to be a big opportunity over the next 10 or 15 years to be playing the role sort of in the repurposing of like I said just really great real estate that's going to need to shift its usage based on some of larger, longer-term trends that we're seeing.

As far as control rights or ownership of land, are the characteristics pretty similar to freestanding multifamily, are there major differences or it depends on the vendor?

I think it depends. I kind of outlined 3 structures there in one -- the first case, that will be analogous to what we did today and basically own and control the residential component. In the second case, it may be again sort of more of a condo structure where we're earning one piece of it, of the sandwich, so to speak, and they're owning the ground plane. In other cases, it's just going to make sense to venture it and there's going to be mutual consent rights as it relates to the asset, whether financing, sale or anything else.

Okay. And then a question on CapEx on our numbers AvalonBay has spent the least amount of CapEx both revenue and maintenance CapEx in the sector. Can you just describe why that's the case? Did you just see better returns on developments or asset recycling? Or is it just a reflection of average age of the portfolio?

Yes. John, I'm happy to address that. If you look at our CapEx, for last year and this year, call it roughly, 5% of NOI -- but if you look at that kind of recurring CapEx. If you look at revenue-enhancing, a fair point, we're spending about $50 a unit. It's dramatically lower than most of our peers. As you may know, we separate our redevelopment bucket and we invest in our assets separate from the same-store pool, which is something that we think is important for investors to understand in terms of how we're allocating that capital and the returns that we generate on that capital. So that certainly has something to do with it. In terms of the nature of the portfolio, certainly, you can go through our portfolio relative to others and look at age and things of that sort. All those factors play into it and we want to make sure that we're being thoughtful about the timing of CapEx, whether it's both defensive and offensive in nature to make sure that we're investing at the appropriate time in the asset and that we're also demonstrating to investors the returns that we're earning by investing in sort of offensive-oriented CapEx. So hopefully, that makes sense, and I hope the questions have been answered.

We'll go next to Drew Babin with Robert W. Baird.

I noticed in the development disclosure quite a bit of progress on the leasing front if you look at projects delivering over the next -- the first units delivering over the next 3 or 4 quarters. At the same time, the average yield on the pipeline ticked down a bit. I'm just wondering if there's anything strategic going on and maybe conceding a little bit on price to increase the velocity of leasing. Or is it something that's just related to things moving in and out of the pipeline.

Drew, this is Sean. I'll take part of that, and then Tim or Matt want to jump in, they're welcome to. In terms of leasing velocity, there's no question that in some markets in particular, to the extent that we're either trying to close out the leasing of the community -- kind of jump-starting it particularly in the third quarter that we will yield a little more on price to get that velocity, either to get the initial lift, as I mentioned, or to finish up, and I would say that was particularly the case as it relates to both North station and Dogpatch this quarter. To give you a sense on Dogpatch as an example, we leased and occupied 32 and 36 a month during the quarter and rent certainly took a little bit of a hit on that one,related to trying to push the velocity before we get into the seasonal slower fourth quarter. In addition on Dogpatch, it's early in the deliveries. We don't have all the amenities. It's just heavy construction. So you tend to yield a little bit more on price during that period of time. So the real rent there we probably won't know until the spring. And then on North station, we're just trying to accelerate the completion of that project in terms of the lease-up. We're in the 70%-plus range there in terms of lease and occupied percentage. We delivered some larger penthouses late in the summer that we're trying to get occupied instead of having them sit there in the winter. Things like that coming into play at asset level, execution plans, and you certainly see that reflected in the rents on some of the assets on the development attachment.

Hey Drew, this is Matt, just as a follow-up to that. As it relates to overall yield, the yield on the development bucket last quarter I think it was a 6.0 and this quarter was 5.8. Probably about half of that a 20 basis point reduction is just changes to the bucket, deals that completed in the second quarter that are not in that bucket any more. They were higher-yielding. And then the other half of it, call it the other 10 basis points, is a little bit of erosion from last quarter in developments that we're leasing both quarters, a lot of that is just the deals this that Sean mentioned.

Great. That's very helpful. And then, I wanted to dig in a little bit on the Lakewood acquisition being that it's kind of the only case study so far in this growth market plan for Denver and Southeast Florida. Specifically in Lakewood, what do you think about that market in terms of barriers to entry, sort of the nimby concept and specific demand drivers that are maybe directly in that market?

Sure. This is Matt. I can take that one as well. It is definitely a submarket that's seen a lot less supply than most of the submarkets in Denver. A lot of the product in Denver is being delivered downtown. This is Lakewood North. It's really on the border between Lakewood and Golden. It's off of I-70 in a master plan kind of controlled community called Denver West, which includes the National Laboratory for Renewable Energy. It includes a fair amount of office. There's a Mills mall across the freeway that was part of it. So the immediate environment is highly controlled by one family that's kind of built it out over the course of decades. But then, even beyond that, when you get out into the broader submarket of which it's apart, Golden is on a permit allocation system, so there's a lot of supply constraints there. And then, Lakewood is its own jurisdiction. It's not as restrictive as Golden but it's certainly not as permissive as Denver. So it is an area that's seeing less supply. And I think, generally, what you see in Denver is the further west you go, the more supply-constrained it gets. And some of that's regulatory and some of that is topographical, you start to get up into the mountain. And this asset is really in the foothills. It's a great place to live. It's for people that are looking to get out on the mountains on a weekend, which is part of that Denver lifestyle, which is so appealing to so many folks. So we were pretty excited about it. And I think, it is a reflection of kind of the locational strategy we're going to take there. At this particular point in the cycle, we certainly -- there's a lot of activity and a lot of excitement about downtown, a lot of people want to live there. There's a lot of new product being delivered. At some point, pricing may make sense to us there, but for right now, we felt that this offered a much better risk-adjusted return.

We'll go next to Wes Golladay with RBC Capital Markets.

Looking at the retail densification opportunity, do you underwrite the developments on a stand-alone basis within the supply end demographics? Or does the quality of the retail asset influence the decision?

Yes, this is Tim here. Well, certainly, the quality of the retail assets, oftentimes, reflective of quality of the location. And so to the extent it's a better location, we're going to demand probably a lower going-in yield, with the expectation we're going to make more of a return on growth over time. But certainly, the existing retail, if it's dying retail and the whole site needs to be sort of redeveloped, that's a very different kind of risk profile. We'd consider it but it's more of a master plan exercise than it is sort surgically where you're going in and maybe pick off part of the parking field or taking down one of the anchors and sort of replanning sort of a lifestyle project that integrates into the existing mall.

Okay. And at what point do you enter these decisions? Are the projects largely entitled? Or do you dictate your skill set and work through the local entitlement process?

We're partnering, in some -- in many cases, with the owner of the real estate. We think, one of the things we offer them is both capital and skill and talent. And so they're leveraging both as part of the relationship. So typically, it's kind of the deals I just described. We'd be sharing in the predevelopment work. We'd be taking most of the lead, as it relates to design and entitlement, but we'll obviously be working alongside with them during that process.

Okay. And then, lastly, you guys talked a lot about volatility of returns and in correlation. Just wondering how you, at the top down, view your portfolio? Do you view it like sort of a sharp ratio? And is that the reason maybe the volatility in Northern California and New York was a little too volatile this quarter and you just wanted to balance it out? Is that why there's reduction in those regions?

No. Not really. I mean, when you look at New York and Northern California, probably just -- we're just -- with respect to Northern California, I think, long-term, we just felt like we're probably a little more over-indexed there at 20% of our existing portfolio relative to the size of the market and our overall portfolio. I mean, Southern California is a much bigger geography. And so even if we were at a market index in Southern California and over-indexed in Northern California, Southern California will still be a bigger part of our portfolio. So it's probably -- we're not necessarily scared by volatility because a lot of times, you can diversify that away through other markets, but we are mindful of when the markets are more volatile is how we invest through the course of the cycle because total returns do matter is when you enter -- when you -- obviously, when you invest the initial capital.

We'll go next to Hardik Goel with Zelman & Associates.

Just on the like-term rent change, you guys did 1.9% in October. And then, if you look at occupancy, it was up 20 basis points, so you could say like-term revenue was up 2.1% in October. What was that compared to last year?

I don't have October last year right in front of me. So I'm happy to get back to you on that offline.

All right. And just one more quick follow-up. When you say your exposure, you're going to take it out in New York, I guess, the target exposure is lower there, where in New York? Because your portfolio is so diversified, and there's a lot of urban, there's a lot of suburban, how have you guys kind of outlined which assets you're going to look to sell and where are those assets?

Maybe a mix of urban and suburban over time. Right now, we're at 24%. We talked about our target being 20%. I don't have the numbers right in front of me but it's sort of roughly split between urban and suburban right now. More recently, we've been selling some of our suburban New York portfolio, in part because that's where we have a deep development pipeline. We've been able to create a lot of value in Northern New Jersey. And then, to some extent, it's strategic. We've been selling out of parts of Connecticut, particularly Northern Fairfield and areas north of that.

And we'll take our next question from John Guinee with Stifel.

Talking a little bit about development. First, hard cost last 12 months, how much have they gone up? And has land adjusted down yet to reflect that increase in hard cost? And then, second, and if you answered this, I apologize, but what yield on cost do you expect in Denver and South Florida, higher yield on cost or lower yield on cost? And what kind of product do you build in, for example, the various South Florida markets, where Miami to West Palm Beach is probably 75, 80 miles distance?

Sure, John. This is Matt. As it relates to the first question, we see hard costs have been growing probably through the course of this year in kind of the mid- to high-single digits, depending on what market you're in. So they're still growing well faster than inflation and faster than rent. So you're right, in the long run, that would impact land valuation. I'm not sure we've seen land values drop yet, but we've certainly seen terms start to soften a little bit in most markets. And land pricing in general is not continuing to go up, maybe with 1 or 2 exceptions. So that's always the first step and it does take a while for the market to adjust, and there's plenty of deals in the pipeline that are out looking for financing. But some of those deals are not penciling the way the sponsors thought they would. So we do expect that there should be continued softening there, but it's certainly -- as we sit here today, it's definitely continuing to get harder to find business that underwrites today because it takes a while for that to play itself out of that process. As it relates to yields and products in the expansion markets, cap rates, I think, in Denver and Southeast Florida are probably pretty similar to what we see, say, in our D.C. markets or in Boston. They're a little higher than the West Coast, they're a little lower than some of the suburban Northeast stuff. So therefore, we would expect we'd be looking for yields probably roughly similar to that, kind of low to mid-6s, depending on the product and the submarket and the location. And our early indications are, certainly, that folks are finding success and finding deals that meet that criteria over time. Yes, Southeast Florida is a very diverse market. That's one of the things we like about it, and, ultimately, over time, just like in our other markets, we would look to have exposure; you look across kind of our portfolio in Metro New York, in Metro D.C., in Metro Boston, we have everything from wood frame garden apartments that are 3 stories in some of the further out suburbs to a lot of kind of mid-rise wrap and podium product that's in infill suburbs, to high-rises downtown, and all those products can be supported in Southeast Florida at different levels. One of the big differences in Florida, obviously, just because of the hurricane, of course none of it is wood frame, it's all concrete, but the costs are very different in these different markets, so you can develop. Certainly, what we've seen is many private developers have been able to develop profitably in that market as well across all those different products types.

John, Tim here. Maybe just a follow-up just briefly here on the land question. Yeah, Matt is right. I mean, the land markets really haven't adjusted. Our focus really has been on one, as it relates to new development, in some cases, densifying our own assets where we've got excess parking, where we could take advantage of it, doing negotiated transactions, either with folks that are looking to repurpose some of their assets. And thirdly, RFPs where we're negotiating basically land residual values with jurisdiction. So those efforts are bearing more fruit than just kind of going to the open market because, in many cases, the land economics still make no sense. I don't think we've tied anything up in Northern California maybe in 3 years from -- in terms of land standpoint. So we're using a lot of different strategies to continue to sort of look at opportunities but most of them are strategies that do allow for some adjustment to the land basis through a negotiated process.

We'll go next to Rich Hill with Morgan Stanley.

So one of the, I guess, more macro trends we're seeing is some declines in bank lending to multifamily. Maybe that's because there's less demand from owners given declining transaction volumes, or maybe that's because of tighter lending standards. But I am wondering if you're seeing any slowdown in supply as a result of declining loan growth? Or is the supply that's in the pipeline pretty much baked at this point?

Yes, Rich, this is Sean. In terms of what's in the pipeline today for '17 and '18, that's pretty much baked, given the duration of the construction cycle. What you're really referring to is deals that might be starting in the fourth quarter or sometime in '18 and delivering in '19 and '20. Certainly, we're hearing about the exact trend you're speaking to as it relates to the tighter lending environment, several of us who were just at ULI just over a week ago heard about that a fair bit from a number of the private developers, both longer-term owners and particularly the merchant builder side, and they're going to their second, third or fourth tier bank to try and get a deal done. So certainly, it's putting more pressure on them in terms of how they're thinking about their deals. And the equity sources that are available to fund the gap because, typically, the deals that are getting done, the loan to cost ratios are coming in a bit lower than earlier in the cycle. So between that, between where construction costs are today and the expectation that based on what we see in the environment, there's potential for construction cost in some market to run above rent growth, it certainly is going to be a regulator on supply, beyond what's already in the pipeline.

Got it. And is there any markets that makes you maybe more bullish or less bullish on?

As it relates to the same-store portfolio or...

Well, really, just overall, in terms of, look, you're a well-capitalized entity, some of the maybe merchant builders are less well-capitalized, so are you seeing opportunities in some markets where maybe some of the smaller competitors are starting to pull back? Or is it too early to see that?

At this point, and Tim can certainly comment as well, but I'd say it's probably a little too early for that. People have their deals capitalized that have already started obviously. The question is what they may have in their pipeline that once they get the permits, once they figure out what the costs are, what the economics are, does it make sense. I'd say, we've probably seen a little bit of that but not a substantial amount at this point. It'd probably take a little bit of time for that to bleed through in terms of opportunities don't get done. And therefore, someone's looking to recycle land. So probably a little early in the cycle for that.

Yes. This is Tim. We have seen -- we started to see some opportunity sort of recycle, maybe land deals we looked at 2 or 3 years ago that maybe less experienced sponsors got into and didn't sort of fully understand the economics of what they're taking on, either in terms of predevelopment requirements or, frankly, construction cost. So you're starting to see a little bit of that. It's happening more in some of the markets I was mentioning earlier, like Northern California or Seattle, which are -- tend to be a little bit more volatile. And maybe there's some folks that are experiencing sort of the wrong side of that volatility at the moment, either in terms of what the numerator or the denominator is telling them when they're running their pro formas.

We'll go next to Alexander Goldfarb with Sandler O'Neill.

Just 2 quick questions here. First, on Southeast Florida, geographically, did you guys outline, are you more concentrated in Miami and the immediate market sort of a little north, a little south? Or are you going all the way up to West Palm, if you could just outline your thoughts?

Yes, this is Matt. We are interested in having a presence in all 3 of the metro areas there, Fort Lauderdale, West Palm and Miami. So it's going to be a function of where transaction volume is, both on the acquisition and development side. But we certainly, over time, would be interested in having presence in all 3 of those submarkets.

Okay. And then, Tim, on your comments on a little deemphasizing New York and San Francisco, how much of that is just a function of the markets on an absolute basis have gotten too expensive, whether it's on cost to develop or acquire, or the absolute rent levels needed to make sense and sort of the other markets may have a broader tenant base and economic base to allow more deals and allow a broader renter profile? So how much of deweighting Northern California in New York is based on pure economics and cost?

I guess, Alex, it's partly that. I mean, part -- in the case of New York, I mean, part of it is just the total size of our overall portfolio. So we're more comfortable with something closer to 20% than 25% as it relates to a weighting, and particularly when you combine it with Boston. In the Northeast, those have been good markets long term but there are some challenges at the same time, whether it's California in the Northeast, from an -- just from a tax base and the like. So part of it is just a little bit of additional diversification, providing other growth opportunities beyond Northeast and California, which, as I said earlier, some of these markets tend to be a little bit more volatile. Timing is important. It's important in New York. It's important in San Francisco. And so, sometimes, you just sort of need to be out of the market. And some of these other markets provide some opportunity to fill in some of the holes as you kind of proceed with your growth strategy.

We'll go next to tie Tayo Okusanya with Jefferies.

Just following up again on the Denver and Southeast Florida strategy. These 2 markets were also the markets that EQR got out of about 2 years ago. I'm just wondering, can you talk a little bit about what their exposure was versus what you're targeting to kind of get a sense if it's the same stuff, whether you're looking at different property type, are you looking at different submarkets? I'm just trying to get a sense of they got out 2 years ago, and now, you guys are getting in, and I'm just trying to understand why.

Yes. I mean, I think, you probably need to talk to EQR to ask them why they got out, but I don't think their portfolios are particularly urban. And I think, they're probably older in those markets. Maybe that played a role. You'd have to ask them. Obviously, they made a bet on more of an urban footprint, so that may have played a role. In our case, we're starting from scratch and so we're looking at trying to build the portfolio that we want to build long term in each of those markets. And as Matt mentioned earlier, I think, the Denver West acquisition was representative of the kind of business we want to do when we think that there's -- when there's maybe an opportunity in a supply-constrained market when there's plenty of supply generally in the market, and there'll be other times when there'll be opportunity in supply-driven markets. So we've said many times, we're agnostic between urban and suburban. We're capable of developing and operating both and we're going to be looking for value at different parts of the cycle where it's most compelling.

Okay. That's helpful. Second question, the installation of the Amazon lockers at some of your locations, could you talk a little bit about just exactly what you're hoping to gain from that? Is it just better management of that process? Is it actually something that ends up dropping to your bottom line? Could you talk about any kind of cost savings that you expect to kind of get from this? And just how much more could you kind of see a rollout of Amazon lockers in your properties?

Yes. Tayo, this is Sean. Just to address this briefly. We did a recent agreement with Amazon to install either their lockers and/or package room across our portfolio, at least a percentage of our portfolio. And if it goes well, we'll probably expand that. I mean, the benefit is multiple. And just to highlight maybe a couple of things, one is just customer convenience that you can pretty much access the package that's been delivered to you 24/7. So regardless of your lifestyle or your work hours, whatever it may be, you can have access to it whenever you want it as opposed to when someone's there. And also, it allows us to probably be a little more efficient in terms of our teams given the number of packages that we receive. There's more than a couple of million packages a year. So it's a pretty significant time allocation in terms of labor hours associated with each one of those packages. Just to give you some simple numbers, we've kind of run through the math, and by the time you receive a package, log in a package, store it, and the customer shows up, you retrieve it, you log it out, you have them sign, et cetera, it can be 3 or 4 minutes of time for each one of those packages, so labor hours definitely adds up. So that's the primary economic benefit associated with the Amazon lockers at this point, and we'll see how it plays out in terms of other opportunities with them as well.

Were you talking about just estimated cost savings something like this could give you, whether it's on a per-building basis or something like that?

Yes, it's a little too early to lay that out. I mean, you pick up some time for someone who's there doing that but they also are doing various other things. So across the portfolio, we certainly expect to pick up something, but we haven't quantified it yet since we're early in the process.

We'll go next to Conor Wagner with Green Street Advisors.

I saw retail bad debt expense picked up in the quarter. Could you speak to a little bit of that? And then, given that you're seeing opportunities to redevelop other retail, could you tell us a little about your strategy for your existing retail?

Conor, it's Sean. I'm happy to address the first one. And then, Matt or Tim may have comments on the second one as well. But on the first one, there's really one tenant that blew out in New York City that we wrote off during the quarter was around $600,000, and that reflects -- this was reflected in the write-off in the comments in the office operations section of the expense attachment is that one write-off.

As it relates to the existing retail, Conor, this is Matt, we have about 600,000 feet, I believe, across our stabilized portfolio and existing retail. And we've actually focused quite a bit in the last couple of years on increasing the occupancy that I think we brought occupancy on that retail up from maybe the low 80s to high 80s around 90% over the last 3 or 4 years. So it's mostly small shop space. So I'm not sure there's any huge opportunities there. But there are select situations. We have a couple of high-profile retail locations where there might be opportunities over time as tenants turn over coming out of some longer-term older leases. But it really is a grab bag, 1,000 feet here, 2,500 feet there. So I don't know if there's any huge opportunities there.

And then, given your current development rights are pretty heavy on New York and New Jersey and your longer-term plan to decrease your allocation in that region, how does that fit? Or how should we think about you're going to backfill the development in the right pipeline?

Conor, this is Matt again. I guess, one of the reasons why you've seen our disposition activity in the last couple of years be a little more heavily concentrated in Metro New York and New Jersey is because of that. So I think, we've sold something like $900 million or $1 billion out of Metro New York the last couple of years, mostly in the suburbs, we did sell Christie Place maybe 3 years ago now. So we are mindful of that in terms of keeping the portfolio in balance. And we have not been doing many acquisitions in that region. So all of our growth there has been through development. So over time, it's just -- it's a big ship. And how you turn it, it gets to kind of the incremental decisions you make on dispositions, but it is still where we're finding tremendous yield in terms of development opportunities, so we're mindful of that and trying to balance that with decisions. Tim, do you want to add something?

Yes. Conor, I think, when we talked about target portfolio, it's in the future. I mean, we are expecting to grow the overall portfolio. So we may not necessarily shrink on an absolute basis the New York Metro area, but we just -- we'll look to recycle capital within that market as we develop. So I think that's kind of the short answer.

Okay. Great. And then, finally, on that same theme, within the Bay Area, do you have any plans to shift your allocation within the 3 regions or hold it pretty constant?

Nothing too specific at the moment, Conor. We're pretty well-distributed between the 3 regions today. So I think, it's going to be probably a little more asset by asset. We're roughly -- of the 20%, we're roughly between 6% and 8% split between Oakland, San Francisco and San Jose today. So we like having a presence in each of those markets. So it's probably going to be probably a bit more opportunistic in terms of where we think sort of the better value from a dispo standpoint sort of presents itself.

(Operator Instructions) We'll go next to Richard Anderson with Mizuho Securities.

Tim, in the beginning of the call, you said you're expanding into other markets, in part just kind of be in business, should there be an opportunity if the cycle turns downward or something along those lines. Can you just describe your mindset there? Are you feeling, just because the number of years is increasing in terms of how long we've been in this cycle, that you have to be a little bit more cautious on the future despite the view on supply maybe coming down next year? Or is that -- am I reading into that too much?

Richard, it's a good question. It's just a recognition that economic cycles don't last forever. Having said that, they don't generally die of old age. They die of distortion. And we're not seeing a lot of distortions today, whether it's in the capital markets or labor markets that suggest we're at the end of the road today. But we have seen -- we do see construction cost and rents that are kind of above trend, and those tend to sort of migrate back towards trend over time. But the one mistake, I think, a lot of developers make is they wait for markets to recover and to see clear evidence that development is profitable before they get in. It oftentimes is 3 or 4 years after you want to be in from a -- particularly from a taking a land position. You kind of need to be feeling with the land markets are like today to be able to sort of understand when the opportunity might be more ripe rather than sort of do it on a hindsight basis. So we're trying to -- it's just kind of what our experience tells us. The best -- oftentimes, the best land deals are either done after there's economic correction, or just before, if it's done on an auction basis where you can oftentimes sort of renegotiate your transaction sort of after the dust has settled.

Okay. And second question is, and this can be a simple yes or no answer, but as a preeminent developer of apartments in the U.S., have you been approached by any cities as a means to sort of sweeten that proposal to attract Amazon's HQ2, like Denver for example? Just kidding about that, but anything along those lines?

There's really nothing there, Rich.

We'll go next to Jeff Pehl with Goldman Sachs.

Just wondering if you could touch on your urban versus suburban rent growth in Northern California and Metro New York, New Jersey and Boston? You had a slide in your 1Q '17 presentation and I'm just wondering if you could give an update on that?

Yes. Jeff, this is Sean. Maybe the way I could answer that for you is tell you a little bit about blended rent change and what you might think about Northern California was one market I heard, just to give you a sense in terms of blended rent change in the third quarter, not a lot different. The East Bay basically was 1.8%, San Francisco was 1.4%. San Jose was only 70 basis points, so that was constrained by the Mountain View rent ordinance that was put into effect. So it's not -- it's got some noise in it in terms of San Jose, so I'd be mindful of that. And I'm sorry, the second market that you indicated was?

Metro New Jersey and New York and then Boston.

Yes. Okay. Boston, I don't have it broken out right in front of me in terms of urban versus suburban rent change during the quarter. Boston overall was in the low 3% range. In terms of New York and New Jersey, that's a little bit easy to break out because it's represented by geography. But to give you a sense, the strongest performing market within our footprint of New York and New Jersey was Long Island where we have rent change around 3%, followed by Central Jersey at about 2.5%, and then, the City of New Jersey about around 2%. The weakest is actually our Westchester portfolio was around 1%, which is a relatively small basket. And we had some movement of 1 or 2 assets that can move the needle so wouldn't read too much into that.

Sean, maybe if you could just talk to just suburban versus urban rent, both East Cost versus West Coast, what we're seeing right now, maybe how that compared versus a couple of quarters ago.

Yes. In terms of urban versus suburban performance, we typically look at the Axiometrics data and then overlay that with our own portfolio data in terms of urban versus suburban performance. One thing just to keep in mind is when we start talking about these kinds of numbers, sample size does matter, and we're about 2/3 suburban. So just keep that in mind, it's not going to be consistent across each and every market in terms of our portfolio allocation between urban and suburban. But across the footprint for our portfolio, urban growth was around 140 basis points versus suburban about 260. So 120 basis point spread in the third quarter as it relates to performance. And again, that's our own portfolio of data as compared to Axiometrics, which is a little bit wider in terms of suburban versus urban, closer to 300 basis points but they have a much bigger sample because they're reporting against as compared to our own portfolio.

And do you think suburban can continue to outperform urban in 2018?

Based on what we know about supply, that would be the expectation. When you look at the deliveries, both '17 and '18, it skewed high on urban submarkets. So if you break apart that, call it, roughly 2% of inventory, it's in the high 1% range for our markets and -- excuse me, for suburban, and then it's also north of 3% for urban. So you blend it together and you get to sort of that 2.1 to 2.3 that I was mentioning earlier. So given that expectation for deliveries, it certainly is good absorption in the urban submarkets but there's just too much supply coming online this year and next year in those environments.

And then, just my second question real quick. The U.S. homeownership rate has ticked up a little bit recently. Just wondering if you have an update across your portfolio on move-outs to purchase a home and if there's any market maybe that are above average for that?

Sure. The move-outs purchase a home still has not moved really this cycle. It's still running around 13% of move-outs. And one thing that we've noticed, and it's been highlighted by a couple of different research firms, is that what we're seeing on the data would indicate that people that have been in some single family rentals are moving into homeownership. And I think, not only for ourselves, but a number of our peers in our markets aren't really seeing the footprint -- or the move-outs to home purchase change much. It sort of makes sense based on what we're seeing. And then, as it relates to any particular markets that are above or below the long-term average, it's typically what you see is the northeast, particularly Boston, is the market that's trending right around or slightly above its long-term average in terms of move-outs to home purchase, which is running right around 22%, 23%. Pretty much all the other markets are at or below the long-term average. To give you a sense, Northern California, it's running around 9% or 10% of the long-term average, closer to 14%. Southern California is pretty much right at its long-term average. So Boston is really the only region where we're seeing that occur at this point.

We'll take our next question from Nick Joseph with Citi.

It's Bilerman. Tim, I was wondering, I know 10 years could be a pretty long time in terms of where the company is going to be, and I think back the last 10 years, you've almost tripled in size. You've gone -- this is pre-Archstone, obviously, with -- from $13 billion of assets to $34 billion today at market. And even post-Archstone, you've grown probably about -- by about 1/3 from call it $25 billion to $34 billion. So as I think about your 10-year forward look in terms of your portfolio construction, how do you think about the size of the portfolio within that context? Could we see that growth continue? Or do you really view this as sort of a steady-state type plan where everything is self-funded, both development as well as growth into new markets via the acquisition and new development?

Yes. Michael, that's a pretty loaded question. I mean, first of all, at scale, we do think there's certain advantages of scale. There's a lot of advantages of scale at the regional level. It's one of the reasons why we have been committed to the footprint that we have been committed to, and in many cases, have exited markets to redeploy our capital into existing markets to develop scale. You can recruit higher-caliber talent into markets when you've got a bigger presence, you can leverage business intelligence in ways that are important, particularly if you're a value-add investor developer in those markets. So we are clearly believers in scale at the regional level. We're a believer at -- in terms of the kind of capabilities that you can build over time, whether it's things like revenue management and other things that you can deploy, where it's not just a matter of just turning a button on. It's supporting that with great market research and data analytics that allow you to sort of optimize it. So all things being equal, we like scale but this is a capital-intensive business. It's only going to make sense to grow the balance sheet at certain points in the cycle. At other times, it may make sense to contract the balance sheet or to recycle. So I would say we have no hard and fast objectives about how big this company should be or ought to be. To some extent, we're a little bit beholden to the capital markets and our cost of capital in terms of how we deploy it, but we do want to achieve certain levels of presence within the markets that we choose to be in, and we wouldn't go into Southeast Florida or Denver, unless we thought we could build a portfolio of a few thousand units over time that would allow us to achieve some of those other strategic advantages.

And it sounds, at least from your initial comments, that doing potential development JVs, approaching acquisitions cautiously is at least some initial ways. I remember back, I can't remember if it was '11 or '12 when you did the exchange with UDR, the SoCal Boston exchange, could that be an avenue where other REITs may be over-indexed -- or under-indexed to the markets you want to get out of, and over-indexed to the ones you want to get into. Is that something that we potentially could see you execute?

It's something we would have interest in, to the extent that there was complementary objectives with another one, whether it's public or private. And then, with respect to development JVs, one thing I didn't mention, we're sponsoring local developers. We're generally -- it'd generally be a structure where we think there's a path to getting sort of full fee ownership of the asset. That's different than maybe kind of a retail JV that we're talking about earlier with respect to GGP, where that really is strategic and where interests are aligned long term and where it might be more of a 50-50 structure on a go-forward basis.

Maybe just last question is on that Slide 13, the conceptual presentation of your circles and then moons, the quarter moons, half moons, just visually, when you look at those items, right, everything Southeast Florida and Denver is below what your legacy markets are. And so, I guess, if these are your market attributes, and even though Southeast Florida and Denver arguably are higher than what you defined as the U.S. market metro average as a half moon, they're still relatively below what your legacy markets are, right? If you did this chart in 10 years and you execute your plan, all your legacy, your new moons, would all have more white.

This is a current assessment of these markets. First of all, if they were more filled in than our existing markets, we would be there already. And so to some extent, you are projecting a little bit too with respect to these markets in terms of how they might perform in the future, for instance, on these sort of knowledge-based economy. Do we think Denver will get more than its fair share of those kinds of jobs? We do, kind of going forward. So -- I mean -- and you look at public investment and infrastructure, I mean, what Denver and the State of Colorado are doing and what's happening in Southeast Florida, it's pretty impressive. Whether it's public transportation or some of the amenities that the public sector is supporting and investing in. So you do -- I mean, we are in a long-term investment business and so we're making -- we are looking at longer-term trends as to where people want to live, the kind of people we rent to as well as what's the posture of the government sector in terms of kind of in terms of supporting this metro area. So we'd like to think that these circles will continue to fill in a bit over time as we get to understand these markets and they continue to mature.

Right. Do you feel like the institutional core capital will come as well, right, you think about your current markets, there's that strong institutional core bid that will allow you to exit...

It's already starting, Michael. It's already investing in these markets. There may not be a big REIT presence in these markets. Honestly, it's one of the reasons we're kind of attracted to them at this particular moment in time where there's not a lot of people that bring the full complement of advantages that we could bring to these markets, whether it's talent or capital. So it kind of presents sort of an interesting opportunity from a competitive standpoint.

And ladies and gentlemen, this does conclude the question-and-answer session. I'd like to turn the call back to Tim Naughton for closing remarks.

Okay. Well, great. Thanks, Aaron. I know it's been a long call but I appreciate you all staying with us. We look forward to seeing you in a -- it's just a couple of weeks time in NAREIT in Dallas. Enjoy the rest of your day.

Ladies and gentlemen, this does conclude today's conference. We thank you for your participation. You may now disconnect.