How to predict veterinary compensation

Most people don't realize that the employment of economic concepts is an innate human behavior. They may not know the terminology, but people have nonetheless engaged in economic behaviors from the beginning of human existence. For example, a Stone Age hunter-gatherer who elected to spend effort to acquire a certain amount of food made choices based on the effort necessary to acquire this food.

To this end, for thousands of years humans collectively hunted the largest animals because they provided the greatest return on investment: considerably less effort was required for a group of 20 to obtain a large 1,000-pound animal than it was for each individual to hunt multiple small animals such as rabbits. Technology (in this case weapons) eventually replaced the labor, and here we are today.

Humans are continuously making choices about how to spend their time to receive optimal benefits. They are constantly revealing their preferences, and their collective actions create market demand. Demand is the willingness of an individual to pay for a specific quantity of a good or service with specific characteristics. This demand provides a relationship between quantity and price-the quantities demanded by individuals in the market and the price they are willing to pay. If the market is robust, then the quantity of goods or services available are all purchased at the price the consumer is willing to pay.

One such relatively robust market is the market for veterinarians. Contrary to popular belief, unemployment in veterinary medicine is less than the natural rate of unemployment. What's more, most veterinarians want to work fewer hours than they currently do. Across the profession, nearly 1,000 more veterinarians would be needed for current members to reduce the hours they work (along with their current compensation) to the level of hours they would optimally prefer. Thus, veterinarians are, on average, fully employed at a wage they are willing to accept in an active market.

As with any average (or mean), there are data points above and below-some veterinarians are earning a lower salary while others are earning more. The most important information is not the average itself but the factors that cause the difference between the mean wage and an individual wage in any given year and the factors that affect year-to-year changes in the average.

First, let's explore differences in income within the same year. (We'll address year-to-year differences in income in the July issue of dvm360.)

Consider the starting salaries of new veterinarians who find employment as companion-animal-exclusive private practitioners in 2014. Table 1 shows the percentages of new veterinarians in this employment type by starting salary. The average starting salary for this practice type was $69,638 in 2014.

You'll notice that there's a wide range of starting salaries: from a low of $24,000 per annum up to a high of $148,000. For companion-animal-exclusive practitioners in 2014, the most frequently reported salary at 13.6 percent of the sample was $70,000 per annum. This was followed by 10.8 percent of the sample reporting anticipated earnings of $65,000 per annum.

While the average salary for a companion-animal-exclusive practitioner in 2014 was $69,638, that does not mean that any individual new veterinarian was actually paid this exact salary. Again, the key question is this: How do we explain why the salaries of individual veterinarians are greater or less than the mean?

For that we turn to statistical analysis, which lets us measure the relationship between starting salary and a number of different factors. And the factors found to explain the difference between mean salary and individual salary are as follows: practice type, additional degrees held, location of job, age, gender, anticipated work hours per week and DVM debt incurred.

Table 2 contains a regression analysis of those factors. (Warning: High-level math ahead. Hang in there for a practical application.) A level of statistical significance less than 0.05 suggests that there is a 95 percent probability that a factor's coefficient (B) is statistically different from zero and, thus, the factor affects starting salaries. The plus or minus sign in front of the factor value indicates whether the factor is associated with a starting income greater than (+) or less than (–) the mean. Summing the (squared) differences between the individual incomes and the mean income is a measure of the total difference (variance). Using all the factors above explains 71.4 percent of this total difference. The remaining difference is likely due to the individual characteristics of each employer or employee.

Using Table 2, we can estimate the expected starting salaries in different situations for new veterinary graduates. For example, suppose I want to know what a 29-year-old female mixed-animal practitioner with $125,000 in DVM debt and a bachelor's degree could, on average, expect to make in the state of New York working 40 hours a week. Starting with the constant (–3311740.914), I would add the value for the year of the survey (1682.893) times the survey year (2014), add the gender adjustment for female (–2438.909), add the practice type adjustment (–4085.009) and add the state adjustment (2591.826). Then I add DVM educational debt (125,000 x 0.008), the product of age (29 x 61.547) and the product of hours worked per week (40 x –124.378). I arrive at a mean starting salary of $71,483.239 for this particular type of practitioner.

Here it is laid out numerically:

Constant –3311740.914 Year of survey +(1682.893 x 2014) Gender adjustment –2438.909 Practice type adjustment –4085.009 State adjustment +2591.826 DVM educational debt +(125,000 x 0.008) Age +(29 x 61.547) Hours worked per week +(40 x –124.378)  

$71,483.239

Notice that we did not add any value corresponding to the fact that this practitioner has a bachelor's degree. Similarly, we would not have added a value for region if computing the expected wage in Region 3 or for a companion-animal-exclusive veterinarian. That's because these and are already accounted for in the constant term, and any other factor not listed in the table is also accounted for in the constant term. (You'll see this table again in future communications from the AVMA.)

So there you have it: the not-so-magical but still-slightly-challenging way to predict veterinary compensation. Using this information, new graduates can make informed decisions about how to maximize their earnings-or accept that certain choices will correspond with a lower income.

Dr. Mike Dicks, director of the AVMA's Veterinary Economics Division, holds a doctorate in agricultural economics from the University of Missouri.