We know that the sodium lamp efficacies (lamp itself not ballast-lamp system) is near 200 lumens/watt for LPS and near 150 lunens/watt for HPS, and even if you add ballast losses that it's about 160 and 120 lm/watt respectively (or is it even less?). That drops to about 130 and 100 lm/watt for typical residential streets for 35-55w LPS & 100w HPS heads.

What is the currently accepted lumen per watt ratings for LED street-lights typically installed for normal residential streets? I'm assuming that high lumen LEDs have higher efficacies just like the sodium lamps.

Streetlights have losses in the optics too. You can generally assume ~.90 efficient optics for a high grade HID lantern

Ballast losses vary a lot with ballast type. Most European chokes are .91+ efficient. Low power (below 100W) ones are a bit less. US autotransformer ballasts (for 250W and up) would probably be a bit less effcient

Efficiency of typical high graqde European HPS lamp (Osram Vialox Super 6Y) :

Rated power (Real power) Lm Eff
50W (52W) 4200Lm 81Lm/W
70W (71W) 6600Lm 93Lm/W
100W (100W) 10700Lm 107Lm/W
150W (152W) 17500Lm 115Lm/W
250W (255W) 33200Lm 129Lm/W
400W (400W) 56500Lm 141Lm/W
600W (600W) 90000Lm 150Lm/W

This means the complete HPS latern efficiency would be about :

Rated power (Real power) Lm Eff
50W (55W) 3780Lm 66Lm/W
70W (77W) 5940Lm 76Lm/W
100W (110W) 9630Lm 88Lm/W
150W (165W) 15750Lm 94Lm/W
250W (275W) 29880Lm 106Lm/W
400W (440W) 50850Lm 115Lm/W
600W (660W) 81000Lm 123Lm/W

LEDs whole system is typically around 100 Lm/W in the better cases, but some are down to 60s

As LEDs are naturally low powe devices, the efficiency with them would theoretically be reverse to power, i.e. lower power lamps more efficient. In reality it depends much more on lantern/cooling design and not really on the total power

Ok interesting.

Generally, are the N.American HPS lamp systems lower than quoted in your tables? By say what percentage?

Also, if an LED maker claims a certain efficacy for their head, is that accurate? Is it "always", "often" or "sometimes" over stated from actual efficacy?

CWA HPS ballasts are very lossy, an S50 250w system will actually draw around 330w (on a quality HPF ballast like Advance). Which is why lower power HPS lamps usually use a simple choke (35w,50w,70w and 100w) to increase the efficiency as the lamp is already less efficient to begin with compared to its larger cousins.

LED is currently all over the map with lumens per watt, ive seen as low as 50 up to 130 in ratings and everything in between. On average however, today's decent quality fixtures are putting out in the high 80's to 90's. What LED has going for it though to boost those numbers is the fact that the   extremely low optical losses in a well designed fixture can more than make up for the raw lumen per watt.

Ok interesting.

Generally, are the N.American HPS lamp systems lower than quoted in your tables? By say what percentage?

The ballast losses of the autotransformer ballasts use to be about double (with isolated output types triple), so whhen European 250W HPS system consumes about 270W, the US is around 290W. But that all is matter of design optimization of cost vs weight vs losses.
On the other hand the fact the lamps have to be operated on some transformer ballast anyway, so it does not matter that much what is the secondary voltage, allows to optimize the lamps for higher efficacy, as there is no limitation on the arc voltage as it is the case for European lamp types intended for just a series reactor ballast. That may gain about 5..10% (depends on the actual power rating, more significant with higher power lamps) back...

Also, if an LED maker claims a certain efficacy for their head, is that accurate? Is it "always", "often" or "sometimes" over stated from actual efficacy?

The standard is clear, 1000lm means 1000lm average across a large population, same as with HID, period.
Other question is, if the published figures are real and really correspond to the given design, or are not based on some "maximize lumen figure" "tuning". But that is the same with all makers, HID's being not different in this matter.
But one difference is there: The rated LED lantern lumen output contains all the losses in the optics, while try to find anything about the optical efficiency with any other lantern - you will find none. So what the 140lm/W means, if you do not know, whether the lantern will eat up 10/10'th or half od the light...
For road light it is common with reputable makers to publish the light distribution charts, but you won't find any information like that for home fixtures.