Why there is an redundant variable debugInit in java.security.AccessControlContext?

Question

I was going through java.security.AccessControlContext, and encountered below piece of code:

private static boolean debugInit = false;
private static Debug debug = null;

static Debug getDebug()
{
    if (debugInit)
        return debug;
    else {
        if (Policy.isSet()) {
            debug = Debug.getInstance("access");
            debugInit = true;
        }
        return debug;
    }
}

Question: why do we need an additional variable debugInit, when we can simple check for if (debug == null) and getInstance() like below:

static Debug getDebug()
{
    if (debug == null && Policy.isSet()) {
        debug = Debug.getInstance("access");
    }
    return debug;
}

Am I missing something or is there any specific reason to use additional variable debugInit (which is not referenced anywhere else) ?