Executive Summary
RevPAR growth rates and employment growth rates over the 1987-2018 period for the Top 25 hotel markets were used to examine the relationship between employment growth and RevPAR growth.
Our findings conclude that while employment growth contributes to RevPAR growth in the short to medium term (e.g. one to five years), its greatest impact is over the long-term (e.g. ten to fifteen years, with ten years the optimum). There appears to be a systematic relationship between RevPAR growth and employment growth over the longer period.
The implications of the report are fourfold. First, employment analysis is relevant to the hotel investment process because employment growth is linked to RevPAR growth in the short-term, but more importantly in the long term (ten to fifteen years). Second, investors should price employment growth into their long-term hotel investments because the long-term relation between employment growth and RevPAR growth is extremely strong. Third, the analysis can be used to rank the MSAs where employment growth is statistically significant in determining RevPAR growth. Fourth, RevPAR growth beta and employment beta are not priced in hotel real estate markets.
Background
Does employment growth generate RevPAR growth? It is generally recognized that the demand for hotel rooms is primarily driven by the general level of economic activity either in the nation or MSA as measured by real metro GDP, real personal income, and employment. But is there a direct link between employment growth and RevPAR growth?
In this brief report we examine the relationship between employment growth of a metropolitan area and its RevPAR performance. Our comparison is motivated by the fact that many hotel investors tend to equate growth with superior RevPAR performance, or at least they tend to justify many hotel investments on the basis of employment growth.
Many investors think that employment growth generates superior RevPAR growth and therefore hotel returns because employment growth tends to result in RevPAR growth and thus rising valuations. According to an analysis by STR, there is a clear relationship between RevPAR and gross operating profit per available room; a 10% increase in revenues equates to a 15%-20% increase in profits (GOPPAR). Employment growth is a demand measure, reflecting only one dimension to the complex market dynamics. Construction determines supply. Demand and supply together drive RevPAR changes which, in turn, influence capital flows that change hotel valuation.
Summary Statistics for the Top 25 Markets Annual Data over the 1987-2018 Period
The table above shows the summary statistics over the 1987-2018 period. Average annual RevPAR growth for each MSA over the thirty-one year period is calculated first. Then this mean RevPAR growth is averaged across twenty-five MSAs. The grand average was 3.6% per annum. The RevPAR figure may appear low, but it takes into account the lodging industry’s slumps in the early 1990s and 2000s. Average RevPAR growth per annum over the thirty-one year period for individual MSAs ranges from a high of 5% to a low of 2.2%. The average employment growth rates across time and markets was a healthy 2.2%. The lowest growth MSA over the thirty-one year period had a growth rate of only 1.3% per annum. The fastest growing MSA in the Top 25 hotel markets, however, was able to grow 3.5% per annum.
RevPAR growth beta for each MSA is calculated by regressing RevPAR growth rates for an MSA against the U.S. Lodging Industry’s RevPAR growth rates over the 1987-2018 period. The betas for individual MSAs range from 0.3 to a high of 1.7, with an average of 1.1. The employment beta, which is the slope of regressing employment growth of an MSA against the U.S. employment growth over the 1987-2018 period, ranges from 0.2 to 1.7. There is a large divergence in RevPAR growth volatility and employment volatility across the MSAs as well.
The Short-run Relationship Between Employment Growth and RevPAR Growth
The question that concerns many hotel investors most is whether employment growth generates RevPAR growth. The accompanying graph shows the cross-sectional time-series plot of RevPAR growth against employment growth. It is apparent from the exhibit that there is a positive relationship between employment growth rates and RevPAR growth rates. Since there are 25 MSAs and each has thirty-one annual observations, there are 775 observations in the graph.
Employment and RevPAR Growth in the Short Term: Cross-Sectional (25 MSA’s) and Time Series Data (31 years)
The regression of RevPAR growth against employment growth has a R2 of 32.2% (32.2 percent of RevPAR growth variation is explained by employment growth variation), with a t-stastic for the slope of 19.163 (highly significant). The results suggest that there is a relationship between employment growth and RevPAR growth in the short-run ( i.e. one year).
Regression Results
The accompanying table shows the results of regressing RevPAR growth against employment growth for each of the Top 25 hotel markets. All except two 25 MSAs have a significant slope, called employment sensitivity, at the 5 percent level. The two MSAs with an insignificant employment sensitivity at the 5 percent level are Philadelphia and New Orleans in order of declining significance.
Results of Regressing RevPAR Growth Against Employment Growth Over 1987-2018
Notes: 1) t-Statistic tests the significance of employment sensitivity. 2) In general, if p-Value is less than 0.05, employment sensitivity is considered statistically significant (at the 5 percent level). 3) R-Squared indicates how much RevPAR growth is explained by employment growth. 4) Correlation is the correlation coefficient between RevPAR growth and employment growth.
The results for the nation and for the pooled cross-sectional time-series data are also shown in the table. The average employment sensitivity for the 25 MSAs was 0.175 with an average t-statistic of 4.44 (these averages not shown in the table). The average R-squared and average correlation coefficient across the 25 markets were, respectively, 0.40 and 0.61. The results suggest that RevPAR growth is positively related to employment growth on a yearly basis.
The table clearly identifies the MSAs where employment growth is statistically significant in determining RevPAR growth. The top ten MSAs with a significant employment sensitivity at the 5 percent level are Orlando, Tampa, Oahu, Phoenix, Seattle, San Francisco, Nashville, Los Angeles, Detroit and Houston in order of increasing significance. We can therefore conclude that, on a yearly basis, RevPAR growth rates are positively linked to employment growth at the metropolitan level as well as nationally.
The accompanying table shows the correlation coefficients between the average RevPAR growth and the average employment growth for a six-year, eight-year, ten-year and twelve-year period. For an intermediate term of six years, the average correlation between RevPAR growth and employment growth for the twenty-six five-year periods was 0.70, indicating that there is a strong positive linkage between RevPAR growth and employment growth. For the twenty-four eight-year periods, this average correlation increased marginally to 0.74. This average further increases to 0.80 for the twenty-two ten-year periods. For the whole twenty twelve year periods, the cross-sectional correlation between RevPAR growth and employment growth declined marginally to 0.78.
Cross-Sectional Correlation Between Average RevPAR Growth and Average Employment Growth Over Different Holding Periods
Private hotel investment is generally considered a medium to long-term investment. As such, these investments are typically evaluated on a seven to ten-year basis, reflecting the long term nature of the investment as well as the convention in the industry. The results in the accompanying table are supportive of a medium to long-term linkage between RevPAR growth and employment growth.
In reality strong employment growth will generate high demand, it is also likely to trigger a supply response. However, over the medium-to-long-term, there is clearly a strong relationship between employment growth and RevPAR growth.
We use either univariate (exponential smoothing & ARIMA), bootstrapping or econometric models to forecast supply, demand, occupancy, ADR, RevPAR, NOI and cap rates for clients at the property, competitive set, market or segment levels with predetermined prediction intervals.
Hotel forecasts tend to rely on the familiar “point” forecast – a single number representing the best estimate of the result. But point forecasts provide no indication of the uncertainity in the number, and uncertainity is an important consideration in decision making.
Call us today to see if your business can benefit from our hotel forecasting services which incorporate prediction intervals in all of our forecasts. Our objective is to generate forecasts as accurate and unbiased as can reasonably be expected – and to do this as efficiently as possible so that our clients can make better decisions.